xk-huang
commited on
Commit
·
86a1106
1
Parent(s):
ffefb5f
[add] model
Browse files- config.json +348 -0
- latest +1 -0
- merges.txt +0 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +6 -0
- tokenizer.json +0 -0
- tokenizer_config.json +9 -0
- trainer_state.json +2194 -0
- training_args.bin +3 -0
- vocab.json +0 -0
- zero_to_fp32.py +587 -0
config.json
ADDED
|
@@ -0,0 +1,348 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_commit_hash": null,
|
| 3 |
+
"_name_or_path": "facebook/sam-vit-huge",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"ScaMultitaskV2Model"
|
| 6 |
+
],
|
| 7 |
+
"cache_dir": "/mnt/blob/weights/.model.cache/",
|
| 8 |
+
"initializer_range": 0.02,
|
| 9 |
+
"mask_caption_decoder_config": {
|
| 10 |
+
"_name_or_path": "",
|
| 11 |
+
"add_cross_attention": false,
|
| 12 |
+
"additional_num_hidden_layers": 12,
|
| 13 |
+
"architectures": null,
|
| 14 |
+
"attention_downsample_rate": 2,
|
| 15 |
+
"bad_words_ids": null,
|
| 16 |
+
"begin_suppress_tokens": null,
|
| 17 |
+
"bos_token_id": null,
|
| 18 |
+
"chunk_size_feed_forward": 0,
|
| 19 |
+
"cross_attention_hidden_size": null,
|
| 20 |
+
"decoder_start_token_id": null,
|
| 21 |
+
"diversity_penalty": 0.0,
|
| 22 |
+
"do_sample": false,
|
| 23 |
+
"early_stopping": false,
|
| 24 |
+
"encoder_no_repeat_ngram_size": 0,
|
| 25 |
+
"eos_token_id": null,
|
| 26 |
+
"exponential_decay_length_penalty": null,
|
| 27 |
+
"finetuning_task": null,
|
| 28 |
+
"forced_bos_token_id": null,
|
| 29 |
+
"forced_eos_token_id": null,
|
| 30 |
+
"hidden_act": "relu",
|
| 31 |
+
"hidden_size": 256,
|
| 32 |
+
"id2label": {
|
| 33 |
+
"0": "LABEL_0",
|
| 34 |
+
"1": "LABEL_1"
|
| 35 |
+
},
|
| 36 |
+
"iou_head_depth": 3,
|
| 37 |
+
"iou_head_hidden_dim": 256,
|
| 38 |
+
"is_decoder": false,
|
| 39 |
+
"is_encoder_decoder": false,
|
| 40 |
+
"label2id": {
|
| 41 |
+
"LABEL_0": 0,
|
| 42 |
+
"LABEL_1": 1
|
| 43 |
+
},
|
| 44 |
+
"layer_norm_eps": 1e-06,
|
| 45 |
+
"length_penalty": 1.0,
|
| 46 |
+
"max_length": 20,
|
| 47 |
+
"min_length": 0,
|
| 48 |
+
"mlp_dim": 2048,
|
| 49 |
+
"model_type": "",
|
| 50 |
+
"no_repeat_ngram_size": 0,
|
| 51 |
+
"num_attention_heads": 8,
|
| 52 |
+
"num_beam_groups": 1,
|
| 53 |
+
"num_beams": 1,
|
| 54 |
+
"num_caption_heads": 1,
|
| 55 |
+
"num_caption_tokens": 8,
|
| 56 |
+
"num_hidden_layers": 2,
|
| 57 |
+
"num_multimask_outputs": 3,
|
| 58 |
+
"num_return_sequences": 1,
|
| 59 |
+
"output_attentions": false,
|
| 60 |
+
"output_hidden_states": false,
|
| 61 |
+
"output_scores": false,
|
| 62 |
+
"pad_token_id": null,
|
| 63 |
+
"prefix": null,
|
| 64 |
+
"problem_type": null,
|
| 65 |
+
"pruned_heads": {},
|
| 66 |
+
"remove_invalid_values": false,
|
| 67 |
+
"repetition_penalty": 1.0,
|
| 68 |
+
"return_dict": true,
|
| 69 |
+
"return_dict_in_generate": false,
|
| 70 |
+
"sep_token_id": null,
|
| 71 |
+
"suppress_tokens": null,
|
| 72 |
+
"task_specific_params": null,
|
| 73 |
+
"temperature": 1.0,
|
| 74 |
+
"tf_legacy_loss": false,
|
| 75 |
+
"tie_encoder_decoder": false,
|
| 76 |
+
"tie_word_embeddings": true,
|
| 77 |
+
"tokenizer_class": null,
|
| 78 |
+
"top_k": 50,
|
| 79 |
+
"top_p": 1.0,
|
| 80 |
+
"torch_dtype": null,
|
| 81 |
+
"torchscript": false,
|
| 82 |
+
"transformers_version": "4.30.2",
|
| 83 |
+
"typical_p": 1.0,
|
| 84 |
+
"use_bfloat16": false
|
| 85 |
+
},
|
| 86 |
+
"model_type": "sca",
|
| 87 |
+
"num_task_tokens": 6,
|
| 88 |
+
"prompt_encoder_config": {
|
| 89 |
+
"_name_or_path": "",
|
| 90 |
+
"add_cross_attention": false,
|
| 91 |
+
"architectures": null,
|
| 92 |
+
"bad_words_ids": null,
|
| 93 |
+
"begin_suppress_tokens": null,
|
| 94 |
+
"bos_token_id": null,
|
| 95 |
+
"chunk_size_feed_forward": 0,
|
| 96 |
+
"cross_attention_hidden_size": null,
|
| 97 |
+
"decoder_start_token_id": null,
|
| 98 |
+
"diversity_penalty": 0.0,
|
| 99 |
+
"do_sample": false,
|
| 100 |
+
"early_stopping": false,
|
| 101 |
+
"encoder_no_repeat_ngram_size": 0,
|
| 102 |
+
"eos_token_id": null,
|
| 103 |
+
"exponential_decay_length_penalty": null,
|
| 104 |
+
"finetuning_task": null,
|
| 105 |
+
"forced_bos_token_id": null,
|
| 106 |
+
"forced_eos_token_id": null,
|
| 107 |
+
"hidden_act": "gelu",
|
| 108 |
+
"hidden_size": 256,
|
| 109 |
+
"id2label": {
|
| 110 |
+
"0": "LABEL_0",
|
| 111 |
+
"1": "LABEL_1"
|
| 112 |
+
},
|
| 113 |
+
"image_embedding_size": 64,
|
| 114 |
+
"image_size": 1024,
|
| 115 |
+
"is_decoder": false,
|
| 116 |
+
"is_encoder_decoder": false,
|
| 117 |
+
"label2id": {
|
| 118 |
+
"LABEL_0": 0,
|
| 119 |
+
"LABEL_1": 1
|
| 120 |
+
},
|
| 121 |
+
"layer_norm_eps": 1e-06,
|
| 122 |
+
"length_penalty": 1.0,
|
| 123 |
+
"mask_input_channels": 16,
|
| 124 |
+
"max_length": 20,
|
| 125 |
+
"min_length": 0,
|
| 126 |
+
"model_type": "",
|
| 127 |
+
"no_repeat_ngram_size": 0,
|
| 128 |
+
"num_beam_groups": 1,
|
| 129 |
+
"num_beams": 1,
|
| 130 |
+
"num_point_embeddings": 4,
|
| 131 |
+
"num_return_sequences": 1,
|
| 132 |
+
"output_attentions": false,
|
| 133 |
+
"output_hidden_states": false,
|
| 134 |
+
"output_scores": false,
|
| 135 |
+
"pad_token_id": null,
|
| 136 |
+
"patch_size": 16,
|
| 137 |
+
"prefix": null,
|
| 138 |
+
"problem_type": null,
|
| 139 |
+
"pruned_heads": {},
|
| 140 |
+
"remove_invalid_values": false,
|
| 141 |
+
"repetition_penalty": 1.0,
|
| 142 |
+
"return_dict": true,
|
| 143 |
+
"return_dict_in_generate": false,
|
| 144 |
+
"sep_token_id": null,
|
| 145 |
+
"suppress_tokens": null,
|
| 146 |
+
"task_specific_params": null,
|
| 147 |
+
"temperature": 1.0,
|
| 148 |
+
"tf_legacy_loss": false,
|
| 149 |
+
"tie_encoder_decoder": false,
|
| 150 |
+
"tie_word_embeddings": true,
|
| 151 |
+
"tokenizer_class": null,
|
| 152 |
+
"top_k": 50,
|
| 153 |
+
"top_p": 1.0,
|
| 154 |
+
"torch_dtype": null,
|
| 155 |
+
"torchscript": false,
|
| 156 |
+
"transformers_version": "4.30.2",
|
| 157 |
+
"typical_p": 1.0,
|
| 158 |
+
"use_bfloat16": false
|
| 159 |
+
},
|
| 160 |
+
"text_config": {
|
| 161 |
+
"_name_or_path": "gpt2-large",
|
| 162 |
+
"activation_function": "gelu_new",
|
| 163 |
+
"add_cross_attention": false,
|
| 164 |
+
"architectures": [
|
| 165 |
+
"GPT2LMHeadModel"
|
| 166 |
+
],
|
| 167 |
+
"attn_pdrop": 0.1,
|
| 168 |
+
"bad_words_ids": null,
|
| 169 |
+
"begin_suppress_tokens": null,
|
| 170 |
+
"bos_token_id": 50256,
|
| 171 |
+
"chunk_size_feed_forward": 0,
|
| 172 |
+
"cross_attention_hidden_size": null,
|
| 173 |
+
"decoder_start_token_id": null,
|
| 174 |
+
"diversity_penalty": 0.0,
|
| 175 |
+
"do_sample": false,
|
| 176 |
+
"early_stopping": false,
|
| 177 |
+
"embd_pdrop": 0.1,
|
| 178 |
+
"encoder_no_repeat_ngram_size": 0,
|
| 179 |
+
"eos_token_id": 50256,
|
| 180 |
+
"exponential_decay_length_penalty": null,
|
| 181 |
+
"finetuning_task": null,
|
| 182 |
+
"forced_bos_token_id": null,
|
| 183 |
+
"forced_eos_token_id": null,
|
| 184 |
+
"id2label": {
|
| 185 |
+
"0": "LABEL_0",
|
| 186 |
+
"1": "LABEL_1"
|
| 187 |
+
},
|
| 188 |
+
"initializer_range": 0.02,
|
| 189 |
+
"is_decoder": false,
|
| 190 |
+
"is_encoder_decoder": false,
|
| 191 |
+
"label2id": {
|
| 192 |
+
"LABEL_0": 0,
|
| 193 |
+
"LABEL_1": 1
|
| 194 |
+
},
|
| 195 |
+
"layer_norm_epsilon": 1e-05,
|
| 196 |
+
"length_penalty": 1.0,
|
| 197 |
+
"max_length": 20,
|
| 198 |
+
"min_length": 0,
|
| 199 |
+
"model_type": "gpt2",
|
| 200 |
+
"n_ctx": 1024,
|
| 201 |
+
"n_embd": 1280,
|
| 202 |
+
"n_head": 20,
|
| 203 |
+
"n_inner": null,
|
| 204 |
+
"n_layer": 36,
|
| 205 |
+
"n_positions": 1024,
|
| 206 |
+
"no_repeat_ngram_size": 0,
|
| 207 |
+
"num_beam_groups": 1,
|
| 208 |
+
"num_beams": 1,
|
| 209 |
+
"num_return_sequences": 1,
|
| 210 |
+
"output_attentions": false,
|
| 211 |
+
"output_hidden_states": false,
|
| 212 |
+
"output_scores": false,
|
| 213 |
+
"pad_token_id": null,
|
| 214 |
+
"prefix": null,
|
| 215 |
+
"problem_type": null,
|
| 216 |
+
"pruned_heads": {},
|
| 217 |
+
"remove_invalid_values": false,
|
| 218 |
+
"reorder_and_upcast_attn": false,
|
| 219 |
+
"repetition_penalty": 1.0,
|
| 220 |
+
"resid_pdrop": 0.1,
|
| 221 |
+
"return_dict": true,
|
| 222 |
+
"return_dict_in_generate": false,
|
| 223 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 224 |
+
"scale_attn_weights": true,
|
| 225 |
+
"sep_token_id": null,
|
| 226 |
+
"summary_activation": null,
|
| 227 |
+
"summary_first_dropout": 0.1,
|
| 228 |
+
"summary_proj_to_labels": true,
|
| 229 |
+
"summary_type": "cls_index",
|
| 230 |
+
"summary_use_proj": true,
|
| 231 |
+
"suppress_tokens": null,
|
| 232 |
+
"task_specific_params": {
|
| 233 |
+
"text-generation": {
|
| 234 |
+
"do_sample": true,
|
| 235 |
+
"max_length": 50
|
| 236 |
+
}
|
| 237 |
+
},
|
| 238 |
+
"temperature": 1.0,
|
| 239 |
+
"tf_legacy_loss": false,
|
| 240 |
+
"tie_encoder_decoder": false,
|
| 241 |
+
"tie_word_embeddings": true,
|
| 242 |
+
"tokenizer_class": null,
|
| 243 |
+
"top_k": 50,
|
| 244 |
+
"top_p": 1.0,
|
| 245 |
+
"torch_dtype": null,
|
| 246 |
+
"torchscript": false,
|
| 247 |
+
"transformers_version": "4.30.2",
|
| 248 |
+
"typical_p": 1.0,
|
| 249 |
+
"use_bfloat16": false,
|
| 250 |
+
"use_cache": true,
|
| 251 |
+
"vocab_size": 50257
|
| 252 |
+
},
|
| 253 |
+
"torch_dtype": "float16",
|
| 254 |
+
"transformers_version": null,
|
| 255 |
+
"use_decoder_only_language_model": true,
|
| 256 |
+
"vision_config": {
|
| 257 |
+
"_name_or_path": "",
|
| 258 |
+
"add_cross_attention": false,
|
| 259 |
+
"architectures": null,
|
| 260 |
+
"attention_dropout": 0.0,
|
| 261 |
+
"bad_words_ids": null,
|
| 262 |
+
"begin_suppress_tokens": null,
|
| 263 |
+
"bos_token_id": null,
|
| 264 |
+
"chunk_size_feed_forward": 0,
|
| 265 |
+
"cross_attention_hidden_size": null,
|
| 266 |
+
"decoder_start_token_id": null,
|
| 267 |
+
"diversity_penalty": 0.0,
|
| 268 |
+
"do_sample": false,
|
| 269 |
+
"dropout": 0.0,
|
| 270 |
+
"early_stopping": false,
|
| 271 |
+
"encoder_no_repeat_ngram_size": 0,
|
| 272 |
+
"eos_token_id": null,
|
| 273 |
+
"exponential_decay_length_penalty": null,
|
| 274 |
+
"finetuning_task": null,
|
| 275 |
+
"forced_bos_token_id": null,
|
| 276 |
+
"forced_eos_token_id": null,
|
| 277 |
+
"global_attn_indexes": [
|
| 278 |
+
7,
|
| 279 |
+
15,
|
| 280 |
+
23,
|
| 281 |
+
31
|
| 282 |
+
],
|
| 283 |
+
"hidden_act": "gelu",
|
| 284 |
+
"hidden_size": 1280,
|
| 285 |
+
"id2label": {
|
| 286 |
+
"0": "LABEL_0",
|
| 287 |
+
"1": "LABEL_1"
|
| 288 |
+
},
|
| 289 |
+
"image_size": 1024,
|
| 290 |
+
"initializer_factor": 1.0,
|
| 291 |
+
"initializer_range": 1e-10,
|
| 292 |
+
"intermediate_size": 6144,
|
| 293 |
+
"is_decoder": false,
|
| 294 |
+
"is_encoder_decoder": false,
|
| 295 |
+
"label2id": {
|
| 296 |
+
"LABEL_0": 0,
|
| 297 |
+
"LABEL_1": 1
|
| 298 |
+
},
|
| 299 |
+
"layer_norm_eps": 1e-06,
|
| 300 |
+
"length_penalty": 1.0,
|
| 301 |
+
"max_length": 20,
|
| 302 |
+
"min_length": 0,
|
| 303 |
+
"mlp_dim": 5120,
|
| 304 |
+
"mlp_ratio": 4.0,
|
| 305 |
+
"model_type": "",
|
| 306 |
+
"no_repeat_ngram_size": 0,
|
| 307 |
+
"num_attention_heads": 16,
|
| 308 |
+
"num_beam_groups": 1,
|
| 309 |
+
"num_beams": 1,
|
| 310 |
+
"num_channels": 3,
|
| 311 |
+
"num_hidden_layers": 32,
|
| 312 |
+
"num_pos_feats": 128,
|
| 313 |
+
"num_return_sequences": 1,
|
| 314 |
+
"output_attentions": false,
|
| 315 |
+
"output_channels": 256,
|
| 316 |
+
"output_hidden_states": false,
|
| 317 |
+
"output_scores": false,
|
| 318 |
+
"pad_token_id": null,
|
| 319 |
+
"patch_size": 16,
|
| 320 |
+
"prefix": null,
|
| 321 |
+
"problem_type": null,
|
| 322 |
+
"projection_dim": 512,
|
| 323 |
+
"pruned_heads": {},
|
| 324 |
+
"qkv_bias": true,
|
| 325 |
+
"remove_invalid_values": false,
|
| 326 |
+
"repetition_penalty": 1.0,
|
| 327 |
+
"return_dict": true,
|
| 328 |
+
"return_dict_in_generate": false,
|
| 329 |
+
"sep_token_id": null,
|
| 330 |
+
"suppress_tokens": null,
|
| 331 |
+
"task_specific_params": null,
|
| 332 |
+
"temperature": 1.0,
|
| 333 |
+
"tf_legacy_loss": false,
|
| 334 |
+
"tie_encoder_decoder": false,
|
| 335 |
+
"tie_word_embeddings": true,
|
| 336 |
+
"tokenizer_class": null,
|
| 337 |
+
"top_k": 50,
|
| 338 |
+
"top_p": 1.0,
|
| 339 |
+
"torch_dtype": null,
|
| 340 |
+
"torchscript": false,
|
| 341 |
+
"transformers_version": "4.30.2",
|
| 342 |
+
"typical_p": 1.0,
|
| 343 |
+
"use_abs_pos": true,
|
| 344 |
+
"use_bfloat16": false,
|
| 345 |
+
"use_rel_pos": true,
|
| 346 |
+
"window_size": 14
|
| 347 |
+
}
|
| 348 |
+
}
|
latest
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
global_step200000
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:42698df7b030f5a734dc22984798482c7ad55a3c843970d16e052b67e678846e
|
| 3 |
+
size 2869532234
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<|endoftext|>",
|
| 3 |
+
"eos_token": "<|endoftext|>",
|
| 4 |
+
"pad_token": "<|endoftext|>",
|
| 5 |
+
"unk_token": "<|endoftext|>"
|
| 6 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"bos_token": "<|endoftext|>",
|
| 4 |
+
"clean_up_tokenization_spaces": true,
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"model_max_length": 20,
|
| 7 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 8 |
+
"unk_token": "<|endoftext|>"
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2194 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 3.5358800888061523,
|
| 3 |
+
"best_model_checkpoint": "/mnt/output/projects/sca-xiaoke-v3/amlt-results/7300886566.81622-380c5407-4540-4a74-b6ec-7def0683f098/checkpoint-185000",
|
| 4 |
+
"epoch": 82.67879288962381,
|
| 5 |
+
"global_step": 200000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"_prepare_inputs_in_ms": 10.839506983757019,
|
| 12 |
+
"compute_loss_in_ms": 1584.6049636602402,
|
| 13 |
+
"epoch": 0.0,
|
| 14 |
+
"learning_rate/full": 0.0,
|
| 15 |
+
"loss": 8.622,
|
| 16 |
+
"step": 1,
|
| 17 |
+
"training_step_in_ms": 2253.9908327162266
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.0,
|
| 21 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 8.927019119262695,
|
| 22 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.023580759441051608,
|
| 23 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 150.659,
|
| 24 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 5.31,
|
| 25 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.166,
|
| 26 |
+
"step": 1
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"_prepare_inputs_in_ms": 4.052688444062369,
|
| 30 |
+
"compute_loss_in_ms": 266.6276198354092,
|
| 31 |
+
"epoch": 0.41,
|
| 32 |
+
"learning_rate/full": 0.0003999851506717227,
|
| 33 |
+
"loss": 4.9777,
|
| 34 |
+
"step": 1000,
|
| 35 |
+
"training_step_in_ms": 963.7469428184035
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"_prepare_inputs_in_ms": 4.001575957983732,
|
| 39 |
+
"compute_loss_in_ms": 266.9940203540027,
|
| 40 |
+
"epoch": 0.83,
|
| 41 |
+
"learning_rate/full": 0.0003999221107895784,
|
| 42 |
+
"loss": 4.4794,
|
| 43 |
+
"step": 2000,
|
| 44 |
+
"training_step_in_ms": 964.067553780973
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"_prepare_inputs_in_ms": 4.017723824828863,
|
| 48 |
+
"compute_loss_in_ms": 266.8530354350805,
|
| 49 |
+
"epoch": 1.24,
|
| 50 |
+
"learning_rate/full": 0.0003998096443163716,
|
| 51 |
+
"loss": 4.3589,
|
| 52 |
+
"step": 3000,
|
| 53 |
+
"training_step_in_ms": 964.8137692287564
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"_prepare_inputs_in_ms": 3.9884973876178265,
|
| 57 |
+
"compute_loss_in_ms": 266.8117158599198,
|
| 58 |
+
"epoch": 1.65,
|
| 59 |
+
"learning_rate/full": 0.0003996477790571026,
|
| 60 |
+
"loss": 4.2668,
|
| 61 |
+
"step": 4000,
|
| 62 |
+
"training_step_in_ms": 963.3689811453223
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"_prepare_inputs_in_ms": 4.001692395657301,
|
| 66 |
+
"compute_loss_in_ms": 267.8194024413824,
|
| 67 |
+
"epoch": 2.07,
|
| 68 |
+
"learning_rate/full": 0.0003994365550295963,
|
| 69 |
+
"loss": 4.2092,
|
| 70 |
+
"step": 5000,
|
| 71 |
+
"training_step_in_ms": 965.0005767121911
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 2.07,
|
| 75 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 4.066258430480957,
|
| 76 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.19624665436714503,
|
| 77 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 81.5743,
|
| 78 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 9.807,
|
| 79 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.306,
|
| 80 |
+
"step": 5000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"_prepare_inputs_in_ms": 4.123009167066434,
|
| 84 |
+
"compute_loss_in_ms": 267.90184961631894,
|
| 85 |
+
"epoch": 2.48,
|
| 86 |
+
"learning_rate/full": 0.0003991760244546079,
|
| 87 |
+
"loss": 4.1535,
|
| 88 |
+
"step": 6000,
|
| 89 |
+
"training_step_in_ms": 963.1004312746227
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"_prepare_inputs_in_ms": 4.096908804029226,
|
| 93 |
+
"compute_loss_in_ms": 268.2109449021518,
|
| 94 |
+
"epoch": 2.89,
|
| 95 |
+
"learning_rate/full": 0.00039886625174291286,
|
| 96 |
+
"loss": 4.1103,
|
| 97 |
+
"step": 7000,
|
| 98 |
+
"training_step_in_ms": 965.0139690972865
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"_prepare_inputs_in_ms": 4.150357250124216,
|
| 102 |
+
"compute_loss_in_ms": 269.2836431860924,
|
| 103 |
+
"epoch": 3.31,
|
| 104 |
+
"learning_rate/full": 0.0003985073134793826,
|
| 105 |
+
"loss": 4.0761,
|
| 106 |
+
"step": 8000,
|
| 107 |
+
"training_step_in_ms": 969.6529387235641
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"_prepare_inputs_in_ms": 4.171766873449087,
|
| 111 |
+
"compute_loss_in_ms": 268.9554896838963,
|
| 112 |
+
"epoch": 3.72,
|
| 113 |
+
"learning_rate/full": 0.00039809973090042857,
|
| 114 |
+
"loss": 4.0582,
|
| 115 |
+
"step": 9000,
|
| 116 |
+
"training_step_in_ms": 964.8072783201933
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"_prepare_inputs_in_ms": 4.219205696135759,
|
| 120 |
+
"compute_loss_in_ms": 268.6923326961696,
|
| 121 |
+
"epoch": 4.13,
|
| 122 |
+
"learning_rate/full": 0.00039764327017710485,
|
| 123 |
+
"loss": 4.0324,
|
| 124 |
+
"step": 10000,
|
| 125 |
+
"training_step_in_ms": 962.4501793310046
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 4.13,
|
| 129 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.913562059402466,
|
| 130 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.22121572228147113,
|
| 131 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 79.9476,
|
| 132 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.007,
|
| 133 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.313,
|
| 134 |
+
"step": 10000
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"_prepare_inputs_in_ms": 4.152017308444512,
|
| 138 |
+
"compute_loss_in_ms": 268.094536613673,
|
| 139 |
+
"epoch": 4.55,
|
| 140 |
+
"learning_rate/full": 0.0003971380439342646,
|
| 141 |
+
"loss": 4.0153,
|
| 142 |
+
"step": 11000,
|
| 143 |
+
"training_step_in_ms": 964.2485933154821
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"_prepare_inputs_in_ms": 4.1431227289140224,
|
| 147 |
+
"compute_loss_in_ms": 268.42261432111263,
|
| 148 |
+
"epoch": 4.96,
|
| 149 |
+
"learning_rate/full": 0.00039658359808291836,
|
| 150 |
+
"loss": 4.0053,
|
| 151 |
+
"step": 12000,
|
| 152 |
+
"training_step_in_ms": 965.7689935192466
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"_prepare_inputs_in_ms": 4.223931197077036,
|
| 156 |
+
"compute_loss_in_ms": 268.6585740670562,
|
| 157 |
+
"epoch": 5.37,
|
| 158 |
+
"learning_rate/full": 0.00039598117829320827,
|
| 159 |
+
"loss": 3.9802,
|
| 160 |
+
"step": 13000,
|
| 161 |
+
"training_step_in_ms": 960.5111146196723
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"_prepare_inputs_in_ms": 4.1840940825641155,
|
| 165 |
+
"compute_loss_in_ms": 268.2862157560885,
|
| 166 |
+
"epoch": 5.79,
|
| 167 |
+
"learning_rate/full": 0.00039533040308028367,
|
| 168 |
+
"loss": 3.9633,
|
| 169 |
+
"step": 14000,
|
| 170 |
+
"training_step_in_ms": 968.1838200092316
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"_prepare_inputs_in_ms": 4.200506843626499,
|
| 174 |
+
"compute_loss_in_ms": 268.91191502287984,
|
| 175 |
+
"epoch": 6.2,
|
| 176 |
+
"learning_rate/full": 0.0003946307092543998,
|
| 177 |
+
"loss": 3.941,
|
| 178 |
+
"step": 15000,
|
| 179 |
+
"training_step_in_ms": 968.3050900287926
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 6.2,
|
| 183 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.8218424320220947,
|
| 184 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2321593168121597,
|
| 185 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.2057,
|
| 186 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.229,
|
| 187 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.32,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"_prepare_inputs_in_ms": 4.21836834491753,
|
| 192 |
+
"compute_loss_in_ms": 268.7001321054995,
|
| 193 |
+
"epoch": 6.61,
|
| 194 |
+
"learning_rate/full": 0.0003938844405523341,
|
| 195 |
+
"loss": 3.9323,
|
| 196 |
+
"step": 16000,
|
| 197 |
+
"training_step_in_ms": 984.531311199069
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"_prepare_inputs_in_ms": 4.2030105367302895,
|
| 201 |
+
"compute_loss_in_ms": 268.3060254715383,
|
| 202 |
+
"epoch": 7.03,
|
| 203 |
+
"learning_rate/full": 0.00039308879047835453,
|
| 204 |
+
"loss": 3.9216,
|
| 205 |
+
"step": 17000,
|
| 206 |
+
"training_step_in_ms": 961.8699175454676
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"_prepare_inputs_in_ms": 4.2327940091490746,
|
| 210 |
+
"compute_loss_in_ms": 269.27323868498206,
|
| 211 |
+
"epoch": 7.44,
|
| 212 |
+
"learning_rate/full": 0.00039224627037346294,
|
| 213 |
+
"loss": 3.8992,
|
| 214 |
+
"step": 18000,
|
| 215 |
+
"training_step_in_ms": 967.632270719856
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"_prepare_inputs_in_ms": 4.229459121823311,
|
| 219 |
+
"compute_loss_in_ms": 268.5666101500392,
|
| 220 |
+
"epoch": 7.85,
|
| 221 |
+
"learning_rate/full": 0.00039135631637799936,
|
| 222 |
+
"loss": 3.8984,
|
| 223 |
+
"step": 19000,
|
| 224 |
+
"training_step_in_ms": 963.238344412297
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"_prepare_inputs_in_ms": 4.240339521318674,
|
| 228 |
+
"compute_loss_in_ms": 268.11582005023956,
|
| 229 |
+
"epoch": 8.27,
|
| 230 |
+
"learning_rate/full": 0.00039041818639024787,
|
| 231 |
+
"loss": 3.8907,
|
| 232 |
+
"step": 20000,
|
| 233 |
+
"training_step_in_ms": 960.5583217255771
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 8.27,
|
| 237 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.7651355266571045,
|
| 238 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2403944760451539,
|
| 239 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 80.0701,
|
| 240 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 9.991,
|
| 241 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.312,
|
| 242 |
+
"step": 20000
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"_prepare_inputs_in_ms": 4.2025597829644274,
|
| 246 |
+
"compute_loss_in_ms": 267.9062583371997,
|
| 247 |
+
"epoch": 8.68,
|
| 248 |
+
"learning_rate/full": 0.00038943398810118026,
|
| 249 |
+
"loss": 3.8736,
|
| 250 |
+
"step": 21000,
|
| 251 |
+
"training_step_in_ms": 962.8509967587888
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"_prepare_inputs_in_ms": 4.1947984509170055,
|
| 255 |
+
"compute_loss_in_ms": 268.05593667179346,
|
| 256 |
+
"epoch": 9.09,
|
| 257 |
+
"learning_rate/full": 0.00038840304981011036,
|
| 258 |
+
"loss": 3.8689,
|
| 259 |
+
"step": 22000,
|
| 260 |
+
"training_step_in_ms": 965.0497910194099
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"_prepare_inputs_in_ms": 4.219467684626579,
|
| 264 |
+
"compute_loss_in_ms": 268.2473221644759,
|
| 265 |
+
"epoch": 9.51,
|
| 266 |
+
"learning_rate/full": 0.0003873256258856351,
|
| 267 |
+
"loss": 3.8553,
|
| 268 |
+
"step": 23000,
|
| 269 |
+
"training_step_in_ms": 967.3452698886395
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"_prepare_inputs_in_ms": 4.203684389591217,
|
| 273 |
+
"compute_loss_in_ms": 268.0240872502327,
|
| 274 |
+
"epoch": 9.92,
|
| 275 |
+
"learning_rate/full": 0.00038620312995292836,
|
| 276 |
+
"loss": 3.8464,
|
| 277 |
+
"step": 24000,
|
| 278 |
+
"training_step_in_ms": 960.33872378245
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"_prepare_inputs_in_ms": 4.218942817300558,
|
| 282 |
+
"compute_loss_in_ms": 268.9947931431234,
|
| 283 |
+
"epoch": 10.33,
|
| 284 |
+
"learning_rate/full": 0.0003850323958934377,
|
| 285 |
+
"loss": 3.8356,
|
| 286 |
+
"step": 25000,
|
| 287 |
+
"training_step_in_ms": 967.6199573352933
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 10.33,
|
| 291 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.71928071975708,
|
| 292 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.24336343080593542,
|
| 293 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 79.8751,
|
| 294 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.016,
|
| 295 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.313,
|
| 296 |
+
"step": 25000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"_prepare_inputs_in_ms": 4.202740984718974,
|
| 300 |
+
"compute_loss_in_ms": 268.5807599723339,
|
| 301 |
+
"epoch": 10.75,
|
| 302 |
+
"learning_rate/full": 0.0003838183948306088,
|
| 303 |
+
"loss": 3.8328,
|
| 304 |
+
"step": 26000,
|
| 305 |
+
"training_step_in_ms": 961.2188336364925
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"_prepare_inputs_in_ms": 4.2269067615270615,
|
| 309 |
+
"compute_loss_in_ms": 267.7264535538852,
|
| 310 |
+
"epoch": 11.16,
|
| 311 |
+
"learning_rate/full": 0.00038255656126572534,
|
| 312 |
+
"loss": 3.8214,
|
| 313 |
+
"step": 27000,
|
| 314 |
+
"training_step_in_ms": 960.6881345175207
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"_prepare_inputs_in_ms": 4.240487921983004,
|
| 318 |
+
"compute_loss_in_ms": 268.31112349405885,
|
| 319 |
+
"epoch": 11.58,
|
| 320 |
+
"learning_rate/full": 0.00038125092378570903,
|
| 321 |
+
"loss": 3.8196,
|
| 322 |
+
"step": 28000,
|
| 323 |
+
"training_step_in_ms": 964.0034716315567
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"_prepare_inputs_in_ms": 4.273211572319269,
|
| 327 |
+
"compute_loss_in_ms": 268.9577633589506,
|
| 328 |
+
"epoch": 11.99,
|
| 329 |
+
"learning_rate/full": 0.0003798991913473777,
|
| 330 |
+
"loss": 3.8171,
|
| 331 |
+
"step": 29000,
|
| 332 |
+
"training_step_in_ms": 973.0124748162925
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"_prepare_inputs_in_ms": 4.244079362601042,
|
| 336 |
+
"compute_loss_in_ms": 267.9022591896355,
|
| 337 |
+
"epoch": 12.4,
|
| 338 |
+
"learning_rate/full": 0.0003785058191464291,
|
| 339 |
+
"loss": 3.8013,
|
| 340 |
+
"step": 30000,
|
| 341 |
+
"training_step_in_ms": 962.8623519428074
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 12.4,
|
| 345 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.687087297439575,
|
| 346 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.24789182257436765,
|
| 347 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.5068,
|
| 348 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.19,
|
| 349 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.318,
|
| 350 |
+
"step": 30000
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"_prepare_inputs_in_ms": 4.183955312502094,
|
| 354 |
+
"compute_loss_in_ms": 267.3523456119001,
|
| 355 |
+
"epoch": 12.82,
|
| 356 |
+
"learning_rate/full": 0.00037706702930065926,
|
| 357 |
+
"loss": 3.8065,
|
| 358 |
+
"step": 31000,
|
| 359 |
+
"training_step_in_ms": 957.8949020504951
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"_prepare_inputs_in_ms": 4.198232229799032,
|
| 363 |
+
"compute_loss_in_ms": 267.4514962993562,
|
| 364 |
+
"epoch": 13.23,
|
| 365 |
+
"learning_rate/full": 0.0003755830450857345,
|
| 366 |
+
"loss": 3.7986,
|
| 367 |
+
"step": 32000,
|
| 368 |
+
"training_step_in_ms": 963.8669461458921
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"_prepare_inputs_in_ms": 4.18659433722496,
|
| 372 |
+
"compute_loss_in_ms": 268.10614936053753,
|
| 373 |
+
"epoch": 13.64,
|
| 374 |
+
"learning_rate/full": 0.0003740572005600189,
|
| 375 |
+
"loss": 3.7923,
|
| 376 |
+
"step": 33000,
|
| 377 |
+
"training_step_in_ms": 965.2572022378445
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"_prepare_inputs_in_ms": 4.225778687745333,
|
| 381 |
+
"compute_loss_in_ms": 268.18950264155865,
|
| 382 |
+
"epoch": 14.06,
|
| 383 |
+
"learning_rate/full": 0.00037248681827609586,
|
| 384 |
+
"loss": 3.7947,
|
| 385 |
+
"step": 34000,
|
| 386 |
+
"training_step_in_ms": 960.6446040645242
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"_prepare_inputs_in_ms": 4.189065434038639,
|
| 390 |
+
"compute_loss_in_ms": 267.6565695255995,
|
| 391 |
+
"epoch": 14.47,
|
| 392 |
+
"learning_rate/full": 0.00037087542640234865,
|
| 393 |
+
"loss": 3.7849,
|
| 394 |
+
"step": 35000,
|
| 395 |
+
"training_step_in_ms": 964.44107465446
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 14.47,
|
| 399 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.6763927936553955,
|
| 400 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.246958162767791,
|
| 401 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.6323,
|
| 402 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.305,
|
| 403 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.322,
|
| 404 |
+
"step": 35000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"_prepare_inputs_in_ms": 4.164576116131573,
|
| 408 |
+
"compute_loss_in_ms": 267.3913585655391,
|
| 409 |
+
"epoch": 14.88,
|
| 410 |
+
"learning_rate/full": 0.0003692218735740006,
|
| 411 |
+
"loss": 3.7838,
|
| 412 |
+
"step": 36000,
|
| 413 |
+
"training_step_in_ms": 965.0988348089159
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"_prepare_inputs_in_ms": 4.195226285606623,
|
| 417 |
+
"compute_loss_in_ms": 268.244338080287,
|
| 418 |
+
"epoch": 15.3,
|
| 419 |
+
"learning_rate/full": 0.00036752484999829976,
|
| 420 |
+
"loss": 3.7779,
|
| 421 |
+
"step": 37000,
|
| 422 |
+
"training_step_in_ms": 966.1850301101804
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"_prepare_inputs_in_ms": 4.189992021769285,
|
| 426 |
+
"compute_loss_in_ms": 267.8583819307387,
|
| 427 |
+
"epoch": 15.71,
|
| 428 |
+
"learning_rate/full": 0.0003657881683678541,
|
| 429 |
+
"loss": 3.7748,
|
| 430 |
+
"step": 38000,
|
| 431 |
+
"training_step_in_ms": 963.9664278812706
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"_prepare_inputs_in_ms": 4.199680101126432,
|
| 435 |
+
"compute_loss_in_ms": 268.06280748173594,
|
| 436 |
+
"epoch": 16.12,
|
| 437 |
+
"learning_rate/full": 0.00036401058098760525,
|
| 438 |
+
"loss": 3.7695,
|
| 439 |
+
"step": 39000,
|
| 440 |
+
"training_step_in_ms": 967.5451415590942
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"_prepare_inputs_in_ms": 4.182811006903648,
|
| 444 |
+
"compute_loss_in_ms": 267.7034317664802,
|
| 445 |
+
"epoch": 16.54,
|
| 446 |
+
"learning_rate/full": 0.00036219252645064074,
|
| 447 |
+
"loss": 3.7688,
|
| 448 |
+
"step": 40000,
|
| 449 |
+
"training_step_in_ms": 967.6426770947874
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 16.54,
|
| 453 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.658348321914673,
|
| 454 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.25104865208221006,
|
| 455 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.4516,
|
| 456 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.329,
|
| 457 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.323,
|
| 458 |
+
"step": 40000
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"_prepare_inputs_in_ms": 4.188235598366435,
|
| 462 |
+
"compute_loss_in_ms": 267.2028581239283,
|
| 463 |
+
"epoch": 16.95,
|
| 464 |
+
"learning_rate/full": 0.0003603344533347134,
|
| 465 |
+
"loss": 3.761,
|
| 466 |
+
"step": 41000,
|
| 467 |
+
"training_step_in_ms": 958.9369925446808
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"_prepare_inputs_in_ms": 4.214274771511555,
|
| 471 |
+
"compute_loss_in_ms": 267.64739087969065,
|
| 472 |
+
"epoch": 17.36,
|
| 473 |
+
"learning_rate/full": 0.00035843490089475537,
|
| 474 |
+
"loss": 3.7527,
|
| 475 |
+
"step": 42000,
|
| 476 |
+
"training_step_in_ms": 966.9140360169113
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"_prepare_inputs_in_ms": 4.22203628346324,
|
| 480 |
+
"compute_loss_in_ms": 268.2027486599982,
|
| 481 |
+
"epoch": 17.78,
|
| 482 |
+
"learning_rate/full": 0.00035650009493379324,
|
| 483 |
+
"loss": 3.7569,
|
| 484 |
+
"step": 43000,
|
| 485 |
+
"training_step_in_ms": 965.1608090028167
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"_prepare_inputs_in_ms": 4.219520688056946,
|
| 489 |
+
"compute_loss_in_ms": 268.0517144687474,
|
| 490 |
+
"epoch": 18.19,
|
| 491 |
+
"learning_rate/full": 0.00035452275921920933,
|
| 492 |
+
"loss": 3.7513,
|
| 493 |
+
"step": 44000,
|
| 494 |
+
"training_step_in_ms": 965.6609862968326
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"_prepare_inputs_in_ms": 4.1790814362466335,
|
| 498 |
+
"compute_loss_in_ms": 267.46442713588476,
|
| 499 |
+
"epoch": 18.6,
|
| 500 |
+
"learning_rate/full": 0.0003525092554160055,
|
| 501 |
+
"loss": 3.7467,
|
| 502 |
+
"step": 45000,
|
| 503 |
+
"training_step_in_ms": 965.7583395838737
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 18.6,
|
| 507 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.6469690799713135,
|
| 508 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.25092950859358,
|
| 509 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.5468,
|
| 510 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.185,
|
| 511 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.318,
|
| 512 |
+
"step": 45000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"_prepare_inputs_in_ms": 4.220321690890847,
|
| 516 |
+
"compute_loss_in_ms": 267.56717593222857,
|
| 517 |
+
"epoch": 19.02,
|
| 518 |
+
"learning_rate/full": 0.0003504601940692656,
|
| 519 |
+
"loss": 3.7497,
|
| 520 |
+
"step": 46000,
|
| 521 |
+
"training_step_in_ms": 964.2354487106204
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"_prepare_inputs_in_ms": 4.205725617706776,
|
| 525 |
+
"compute_loss_in_ms": 268.48560455814004,
|
| 526 |
+
"epoch": 19.43,
|
| 527 |
+
"learning_rate/full": 0.0003483698657658315,
|
| 528 |
+
"loss": 3.7388,
|
| 529 |
+
"step": 47000,
|
| 530 |
+
"training_step_in_ms": 967.1772802136838
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"_prepare_inputs_in_ms": 4.253620557487011,
|
| 534 |
+
"compute_loss_in_ms": 267.8220782019198,
|
| 535 |
+
"epoch": 19.84,
|
| 536 |
+
"learning_rate/full": 0.0003462450012513184,
|
| 537 |
+
"loss": 3.744,
|
| 538 |
+
"step": 48000,
|
| 539 |
+
"training_step_in_ms": 962.239847779274
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"_prepare_inputs_in_ms": 4.241022698581219,
|
| 543 |
+
"compute_loss_in_ms": 268.39736769348383,
|
| 544 |
+
"epoch": 20.26,
|
| 545 |
+
"learning_rate/full": 0.00034408405297105753,
|
| 546 |
+
"loss": 3.7372,
|
| 547 |
+
"step": 49000,
|
| 548 |
+
"training_step_in_ms": 966.1254425011575
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"_prepare_inputs_in_ms": 4.21300457790494,
|
| 552 |
+
"compute_loss_in_ms": 268.7097208276391,
|
| 553 |
+
"epoch": 20.67,
|
| 554 |
+
"learning_rate/full": 0.0003418853377786221,
|
| 555 |
+
"loss": 3.7334,
|
| 556 |
+
"step": 50000,
|
| 557 |
+
"training_step_in_ms": 968.8591329194605
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 20.67,
|
| 561 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.632450580596924,
|
| 562 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.253658040934189,
|
| 563 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.9084,
|
| 564 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.268,
|
| 565 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.321,
|
| 566 |
+
"step": 50000
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"_prepare_inputs_in_ms": 4.228368219805927,
|
| 570 |
+
"compute_loss_in_ms": 268.4243040457368,
|
| 571 |
+
"epoch": 21.08,
|
| 572 |
+
"learning_rate/full": 0.0003396560466114797,
|
| 573 |
+
"loss": 3.7438,
|
| 574 |
+
"step": 51000,
|
| 575 |
+
"training_step_in_ms": 964.8838895820081
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"_prepare_inputs_in_ms": 4.233973186463118,
|
| 579 |
+
"compute_loss_in_ms": 268.22105176746845,
|
| 580 |
+
"epoch": 21.5,
|
| 581 |
+
"learning_rate/full": 0.0003373900810764743,
|
| 582 |
+
"loss": 3.728,
|
| 583 |
+
"step": 52000,
|
| 584 |
+
"training_step_in_ms": 962.5837270207703
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"_prepare_inputs_in_ms": 4.230040587484837,
|
| 588 |
+
"compute_loss_in_ms": 267.9561189264059,
|
| 589 |
+
"epoch": 21.91,
|
| 590 |
+
"learning_rate/full": 0.0003350878976336386,
|
| 591 |
+
"loss": 3.7287,
|
| 592 |
+
"step": 53000,
|
| 593 |
+
"training_step_in_ms": 964.1780665256083
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"_prepare_inputs_in_ms": 4.2476331405341625,
|
| 597 |
+
"compute_loss_in_ms": 268.6795903816819,
|
| 598 |
+
"epoch": 22.32,
|
| 599 |
+
"learning_rate/full": 0.0003327546685845955,
|
| 600 |
+
"loss": 3.7223,
|
| 601 |
+
"step": 54000,
|
| 602 |
+
"training_step_in_ms": 967.9701336547732
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"_prepare_inputs_in_ms": 4.220093585550785,
|
| 606 |
+
"compute_loss_in_ms": 267.76603213325143,
|
| 607 |
+
"epoch": 22.74,
|
| 608 |
+
"learning_rate/full": 0.00033038868430752995,
|
| 609 |
+
"loss": 3.7261,
|
| 610 |
+
"step": 55000,
|
| 611 |
+
"training_step_in_ms": 962.9226383566856
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 22.74,
|
| 615 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.6242456436157227,
|
| 616 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.25521777862203787,
|
| 617 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.5774,
|
| 618 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.181,
|
| 619 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.318,
|
| 620 |
+
"step": 55000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"_prepare_inputs_in_ms": 4.227604019205745,
|
| 624 |
+
"compute_loss_in_ms": 267.647510971874,
|
| 625 |
+
"epoch": 23.15,
|
| 626 |
+
"learning_rate/full": 0.00032799052857365924,
|
| 627 |
+
"loss": 3.7232,
|
| 628 |
+
"step": 56000,
|
| 629 |
+
"training_step_in_ms": 963.1331409327686
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"_prepare_inputs_in_ms": 4.1824629083275795,
|
| 633 |
+
"compute_loss_in_ms": 267.5452450104058,
|
| 634 |
+
"epoch": 23.56,
|
| 635 |
+
"learning_rate/full": 0.0003255583453025672,
|
| 636 |
+
"loss": 3.7153,
|
| 637 |
+
"step": 57000,
|
| 638 |
+
"training_step_in_ms": 965.8048706538975
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"_prepare_inputs_in_ms": 4.190924048423767,
|
| 642 |
+
"compute_loss_in_ms": 268.4611966535449,
|
| 643 |
+
"epoch": 23.98,
|
| 644 |
+
"learning_rate/full": 0.0003230975988657048,
|
| 645 |
+
"loss": 3.72,
|
| 646 |
+
"step": 58000,
|
| 647 |
+
"training_step_in_ms": 965.0302759557962
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"_prepare_inputs_in_ms": 4.20093110203743,
|
| 651 |
+
"compute_loss_in_ms": 267.884086355567,
|
| 652 |
+
"epoch": 24.39,
|
| 653 |
+
"learning_rate/full": 0.0003206064799382713,
|
| 654 |
+
"loss": 3.7127,
|
| 655 |
+
"step": 59000,
|
| 656 |
+
"training_step_in_ms": 964.0987507812679
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"_prepare_inputs_in_ms": 4.210771672427654,
|
| 660 |
+
"compute_loss_in_ms": 267.6700286902487,
|
| 661 |
+
"epoch": 24.8,
|
| 662 |
+
"learning_rate/full": 0.00031808560316658635,
|
| 663 |
+
"loss": 3.7115,
|
| 664 |
+
"step": 60000,
|
| 665 |
+
"training_step_in_ms": 965.6994955539703
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 24.8,
|
| 669 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.61262583732605,
|
| 670 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.259026039077661,
|
| 671 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.8405,
|
| 672 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.277,
|
| 673 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.321,
|
| 674 |
+
"step": 60000
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"_prepare_inputs_in_ms": 4.141795973225338,
|
| 678 |
+
"compute_loss_in_ms": 267.2422932982445,
|
| 679 |
+
"epoch": 25.22,
|
| 680 |
+
"learning_rate/full": 0.0003155330235866319,
|
| 681 |
+
"loss": 3.7085,
|
| 682 |
+
"step": 61000,
|
| 683 |
+
"training_step_in_ms": 965.9709356427193
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"_prepare_inputs_in_ms": 4.12061008810997,
|
| 687 |
+
"compute_loss_in_ms": 268.04699283093214,
|
| 688 |
+
"epoch": 25.63,
|
| 689 |
+
"learning_rate/full": 0.000312954476063518,
|
| 690 |
+
"loss": 3.7031,
|
| 691 |
+
"step": 62000,
|
| 692 |
+
"training_step_in_ms": 964.8432326950133
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"_prepare_inputs_in_ms": 4.14773878082633,
|
| 696 |
+
"compute_loss_in_ms": 268.0800936706364,
|
| 697 |
+
"epoch": 26.04,
|
| 698 |
+
"learning_rate/full": 0.00031035068146119334,
|
| 699 |
+
"loss": 3.7053,
|
| 700 |
+
"step": 63000,
|
| 701 |
+
"training_step_in_ms": 964.9066540151834
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"_prepare_inputs_in_ms": 4.157721221446991,
|
| 705 |
+
"compute_loss_in_ms": 267.8991154767573,
|
| 706 |
+
"epoch": 26.46,
|
| 707 |
+
"learning_rate/full": 0.00030771441463138695,
|
| 708 |
+
"loss": 3.6991,
|
| 709 |
+
"step": 64000,
|
| 710 |
+
"training_step_in_ms": 966.7548437044024
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"_prepare_inputs_in_ms": 4.127573002129793,
|
| 714 |
+
"compute_loss_in_ms": 268.0769842043519,
|
| 715 |
+
"epoch": 26.87,
|
| 716 |
+
"learning_rate/full": 0.00030505419362911944,
|
| 717 |
+
"loss": 3.7026,
|
| 718 |
+
"step": 65000,
|
| 719 |
+
"training_step_in_ms": 966.6518254801631
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 26.87,
|
| 723 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.604250907897949,
|
| 724 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2600293381296915,
|
| 725 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.2046,
|
| 726 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.362,
|
| 727 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.324,
|
| 728 |
+
"step": 65000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"_prepare_inputs_in_ms": 4.138722216210714,
|
| 732 |
+
"compute_loss_in_ms": 268.0730670392513,
|
| 733 |
+
"epoch": 27.28,
|
| 734 |
+
"learning_rate/full": 0.0003023653504808654,
|
| 735 |
+
"loss": 3.6903,
|
| 736 |
+
"step": 66000,
|
| 737 |
+
"training_step_in_ms": 965.3969647027552
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"_prepare_inputs_in_ms": 4.153087247163057,
|
| 741 |
+
"compute_loss_in_ms": 268.1363028138876,
|
| 742 |
+
"epoch": 27.7,
|
| 743 |
+
"learning_rate/full": 0.0002996566527388639,
|
| 744 |
+
"loss": 3.6969,
|
| 745 |
+
"step": 67000,
|
| 746 |
+
"training_step_in_ms": 965.2745163962245
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"_prepare_inputs_in_ms": 4.149796262383461,
|
| 750 |
+
"compute_loss_in_ms": 267.83912086486816,
|
| 751 |
+
"epoch": 28.11,
|
| 752 |
+
"learning_rate/full": 0.0002969179137925403,
|
| 753 |
+
"loss": 3.6937,
|
| 754 |
+
"step": 68000,
|
| 755 |
+
"training_step_in_ms": 968.0436515249312
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"_prepare_inputs_in_ms": 4.1318743117153645,
|
| 759 |
+
"compute_loss_in_ms": 268.5812944062054,
|
| 760 |
+
"epoch": 28.52,
|
| 761 |
+
"learning_rate/full": 0.0002941579883457959,
|
| 762 |
+
"loss": 3.6891,
|
| 763 |
+
"step": 69000,
|
| 764 |
+
"training_step_in_ms": 964.7220857255161
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"_prepare_inputs_in_ms": 4.13500452041626,
|
| 768 |
+
"compute_loss_in_ms": 267.81502260267735,
|
| 769 |
+
"epoch": 28.94,
|
| 770 |
+
"learning_rate/full": 0.0002913748308243434,
|
| 771 |
+
"loss": 3.6864,
|
| 772 |
+
"step": 70000,
|
| 773 |
+
"training_step_in_ms": 964.9831298328936
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 28.94,
|
| 777 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5996363162994385,
|
| 778 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.26106738202465474,
|
| 779 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 79.1905,
|
| 780 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.102,
|
| 781 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.316,
|
| 782 |
+
"step": 70000
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"_prepare_inputs_in_ms": 4.127892008641871,
|
| 786 |
+
"compute_loss_in_ms": 267.3319378942251,
|
| 787 |
+
"epoch": 29.35,
|
| 788 |
+
"learning_rate/full": 0.000288569127930656,
|
| 789 |
+
"loss": 3.6864,
|
| 790 |
+
"step": 71000,
|
| 791 |
+
"training_step_in_ms": 964.4970440678298
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"_prepare_inputs_in_ms": 4.127237547188997,
|
| 795 |
+
"compute_loss_in_ms": 267.12262638285756,
|
| 796 |
+
"epoch": 29.76,
|
| 797 |
+
"learning_rate/full": 0.0002857387308278068,
|
| 798 |
+
"loss": 3.688,
|
| 799 |
+
"step": 72000,
|
| 800 |
+
"training_step_in_ms": 963.8779099695385
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"_prepare_inputs_in_ms": 4.134644038975239,
|
| 804 |
+
"compute_loss_in_ms": 268.62900394946337,
|
| 805 |
+
"epoch": 30.18,
|
| 806 |
+
"learning_rate/full": 0.0002828899985518552,
|
| 807 |
+
"loss": 3.6777,
|
| 808 |
+
"step": 73000,
|
| 809 |
+
"training_step_in_ms": 966.1114624030888
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"_prepare_inputs_in_ms": 4.121612053364515,
|
| 813 |
+
"compute_loss_in_ms": 267.61186150833964,
|
| 814 |
+
"epoch": 30.59,
|
| 815 |
+
"learning_rate/full": 0.00028002081440907064,
|
| 816 |
+
"loss": 3.6805,
|
| 817 |
+
"step": 74000,
|
| 818 |
+
"training_step_in_ms": 966.5145794674754
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"_prepare_inputs_in_ms": 4.143208879977465,
|
| 822 |
+
"compute_loss_in_ms": 268.1201763525605,
|
| 823 |
+
"epoch": 31.0,
|
| 824 |
+
"learning_rate/full": 0.00027713188632770775,
|
| 825 |
+
"loss": 3.6879,
|
| 826 |
+
"step": 75000,
|
| 827 |
+
"training_step_in_ms": 965.0215070433915
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 31.0,
|
| 831 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5912275314331055,
|
| 832 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.26092454356630995,
|
| 833 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.6905,
|
| 834 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.166,
|
| 835 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.318,
|
| 836 |
+
"step": 75000
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"_prepare_inputs_in_ms": 4.137767382511279,
|
| 840 |
+
"compute_loss_in_ms": 267.779229991138,
|
| 841 |
+
"epoch": 31.42,
|
| 842 |
+
"learning_rate/full": 0.00027422684725453034,
|
| 843 |
+
"loss": 3.6725,
|
| 844 |
+
"step": 76000,
|
| 845 |
+
"training_step_in_ms": 963.9590919055045
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"_prepare_inputs_in_ms": 4.138436369597912,
|
| 849 |
+
"compute_loss_in_ms": 268.5314156524837,
|
| 850 |
+
"epoch": 31.83,
|
| 851 |
+
"learning_rate/full": 0.0002712976542440004,
|
| 852 |
+
"loss": 3.6758,
|
| 853 |
+
"step": 77000,
|
| 854 |
+
"training_step_in_ms": 964.5805881880224
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"_prepare_inputs_in_ms": 4.146489001810551,
|
| 858 |
+
"compute_loss_in_ms": 268.08270847052336,
|
| 859 |
+
"epoch": 32.24,
|
| 860 |
+
"learning_rate/full": 0.00026835083436875734,
|
| 861 |
+
"loss": 3.6692,
|
| 862 |
+
"step": 78000,
|
| 863 |
+
"training_step_in_ms": 966.2784307040274
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"_prepare_inputs_in_ms": 4.1479257568717,
|
| 867 |
+
"compute_loss_in_ms": 267.2712250612676,
|
| 868 |
+
"epoch": 32.66,
|
| 869 |
+
"learning_rate/full": 0.0002653930599834242,
|
| 870 |
+
"loss": 3.6736,
|
| 871 |
+
"step": 79000,
|
| 872 |
+
"training_step_in_ms": 968.4041320718825
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"_prepare_inputs_in_ms": 4.126915082335472,
|
| 876 |
+
"compute_loss_in_ms": 268.5138017758727,
|
| 877 |
+
"epoch": 33.07,
|
| 878 |
+
"learning_rate/full": 0.0002624132077738845,
|
| 879 |
+
"loss": 3.6731,
|
| 880 |
+
"step": 80000,
|
| 881 |
+
"training_step_in_ms": 969.1153637133539
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 33.07,
|
| 885 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5850701332092285,
|
| 886 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2637883107097123,
|
| 887 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.0939,
|
| 888 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.377,
|
| 889 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.324,
|
| 890 |
+
"step": 80000
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"_prepare_inputs_in_ms": 4.120095092348937,
|
| 894 |
+
"compute_loss_in_ms": 267.7357228696346,
|
| 895 |
+
"epoch": 33.48,
|
| 896 |
+
"learning_rate/full": 0.0002594179251945605,
|
| 897 |
+
"loss": 3.6671,
|
| 898 |
+
"step": 81000,
|
| 899 |
+
"training_step_in_ms": 963.6873134560883
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"_prepare_inputs_in_ms": 4.124605063349009,
|
| 903 |
+
"compute_loss_in_ms": 267.57319816574454,
|
| 904 |
+
"epoch": 33.9,
|
| 905 |
+
"learning_rate/full": 0.00025641096982950234,
|
| 906 |
+
"loss": 3.6662,
|
| 907 |
+
"step": 82000,
|
| 908 |
+
"training_step_in_ms": 965.3064449094236
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"_prepare_inputs_in_ms": 4.129976071417332,
|
| 912 |
+
"compute_loss_in_ms": 267.9822950810194,
|
| 913 |
+
"epoch": 34.31,
|
| 914 |
+
"learning_rate/full": 0.00025339009590173424,
|
| 915 |
+
"loss": 3.6651,
|
| 916 |
+
"step": 83000,
|
| 917 |
+
"training_step_in_ms": 967.112907551229
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"_prepare_inputs_in_ms": 4.144246697425842,
|
| 921 |
+
"compute_loss_in_ms": 268.09172417223454,
|
| 922 |
+
"epoch": 34.73,
|
| 923 |
+
"learning_rate/full": 0.0002503530053325778,
|
| 924 |
+
"loss": 3.6622,
|
| 925 |
+
"step": 84000,
|
| 926 |
+
"training_step_in_ms": 963.3224161304533
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"_prepare_inputs_in_ms": 4.138608705252409,
|
| 930 |
+
"compute_loss_in_ms": 267.34737430512905,
|
| 931 |
+
"epoch": 35.14,
|
| 932 |
+
"learning_rate/full": 0.0002473065215350535,
|
| 933 |
+
"loss": 3.66,
|
| 934 |
+
"step": 85000,
|
| 935 |
+
"training_step_in_ms": 964.2706917002797
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 35.14,
|
| 939 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5811269283294678,
|
| 940 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2639846881301574,
|
| 941 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.4776,
|
| 942 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.194,
|
| 943 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.319,
|
| 944 |
+
"step": 85000
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"_prepare_inputs_in_ms": 4.154233903419681,
|
| 948 |
+
"compute_loss_in_ms": 267.71352230757475,
|
| 949 |
+
"epoch": 35.55,
|
| 950 |
+
"learning_rate/full": 0.00024424836556120313,
|
| 951 |
+
"loss": 3.6547,
|
| 952 |
+
"step": 86000,
|
| 953 |
+
"training_step_in_ms": 962.3722572363913
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"_prepare_inputs_in_ms": 4.109987150877714,
|
| 957 |
+
"compute_loss_in_ms": 267.4717643670738,
|
| 958 |
+
"epoch": 35.97,
|
| 959 |
+
"learning_rate/full": 0.000241179291965253,
|
| 960 |
+
"loss": 3.662,
|
| 961 |
+
"step": 87000,
|
| 962 |
+
"training_step_in_ms": 962.7631023935974
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"_prepare_inputs_in_ms": 4.130292858928442,
|
| 966 |
+
"compute_loss_in_ms": 268.7110885903239,
|
| 967 |
+
"epoch": 36.38,
|
| 968 |
+
"learning_rate/full": 0.0002381000579951894,
|
| 969 |
+
"loss": 3.6533,
|
| 970 |
+
"step": 88000,
|
| 971 |
+
"training_step_in_ms": 966.5117364116013
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"_prepare_inputs_in_ms": 4.127315446734428,
|
| 975 |
+
"compute_loss_in_ms": 267.80369279161096,
|
| 976 |
+
"epoch": 36.79,
|
| 977 |
+
"learning_rate/full": 0.00023501142340591894,
|
| 978 |
+
"loss": 3.6541,
|
| 979 |
+
"step": 89000,
|
| 980 |
+
"training_step_in_ms": 966.1625612042844
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"_prepare_inputs_in_ms": 4.147790517657995,
|
| 984 |
+
"compute_loss_in_ms": 268.44236666709185,
|
| 985 |
+
"epoch": 37.21,
|
| 986 |
+
"learning_rate/full": 0.00023191415027181022,
|
| 987 |
+
"loss": 3.6505,
|
| 988 |
+
"step": 90000,
|
| 989 |
+
"training_step_in_ms": 969.9625728055835
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 37.21,
|
| 993 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5739023685455322,
|
| 994 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2666824586460844,
|
| 995 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.5693,
|
| 996 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.182,
|
| 997 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.318,
|
| 998 |
+
"step": 90000
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"_prepare_inputs_in_ms": 4.1041579188370125,
|
| 1002 |
+
"compute_loss_in_ms": 267.84072072431445,
|
| 1003 |
+
"epoch": 37.62,
|
| 1004 |
+
"learning_rate/full": 0.00022880900279866363,
|
| 1005 |
+
"loss": 3.6517,
|
| 1006 |
+
"step": 91000,
|
| 1007 |
+
"training_step_in_ms": 962.2163318134844
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"_prepare_inputs_in_ms": 4.135395355522633,
|
| 1011 |
+
"compute_loss_in_ms": 267.92124405503273,
|
| 1012 |
+
"epoch": 38.03,
|
| 1013 |
+
"learning_rate/full": 0.0002256936284593779,
|
| 1014 |
+
"loss": 3.6526,
|
| 1015 |
+
"step": 92000,
|
| 1016 |
+
"training_step_in_ms": 965.9340194314718
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"_prepare_inputs_in_ms": 4.124495934695005,
|
| 1020 |
+
"compute_loss_in_ms": 267.63855477049947,
|
| 1021 |
+
"epoch": 38.45,
|
| 1022 |
+
"learning_rate/full": 0.00022257502654664658,
|
| 1023 |
+
"loss": 3.6411,
|
| 1024 |
+
"step": 93000,
|
| 1025 |
+
"training_step_in_ms": 968.471509065479
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"_prepare_inputs_in_ms": 4.13615457713604,
|
| 1029 |
+
"compute_loss_in_ms": 268.6791280247271,
|
| 1030 |
+
"epoch": 38.86,
|
| 1031 |
+
"learning_rate/full": 0.00021945398441148287,
|
| 1032 |
+
"loss": 3.6456,
|
| 1033 |
+
"step": 94000,
|
| 1034 |
+
"training_step_in_ms": 964.7193784303963
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"_prepare_inputs_in_ms": 4.138962186872959,
|
| 1038 |
+
"compute_loss_in_ms": 267.87416788190603,
|
| 1039 |
+
"epoch": 39.27,
|
| 1040 |
+
"learning_rate/full": 0.00021632188341385878,
|
| 1041 |
+
"loss": 3.6474,
|
| 1042 |
+
"step": 95000,
|
| 1043 |
+
"training_step_in_ms": 964.3936127200723
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 39.27,
|
| 1047 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.570530414581299,
|
| 1048 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.26555671498307476,
|
| 1049 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.8889,
|
| 1050 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.271,
|
| 1051 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.321,
|
| 1052 |
+
"step": 95000
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"_prepare_inputs_in_ms": 4.114193618297577,
|
| 1056 |
+
"compute_loss_in_ms": 267.61803087219596,
|
| 1057 |
+
"epoch": 39.69,
|
| 1058 |
+
"learning_rate/full": 0.0002131888850633025,
|
| 1059 |
+
"loss": 3.6443,
|
| 1060 |
+
"step": 96000,
|
| 1061 |
+
"training_step_in_ms": 962.9699364975095
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"_prepare_inputs_in_ms": 4.118229653686285,
|
| 1065 |
+
"compute_loss_in_ms": 267.60003339126706,
|
| 1066 |
+
"epoch": 40.1,
|
| 1067 |
+
"learning_rate/full": 0.00021005263255270636,
|
| 1068 |
+
"loss": 3.6436,
|
| 1069 |
+
"step": 97000,
|
| 1070 |
+
"training_step_in_ms": 967.0044349320233
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"_prepare_inputs_in_ms": 4.14753895252943,
|
| 1074 |
+
"compute_loss_in_ms": 267.94382878765464,
|
| 1075 |
+
"epoch": 40.51,
|
| 1076 |
+
"learning_rate/full": 0.00020691389970544863,
|
| 1077 |
+
"loss": 3.6409,
|
| 1078 |
+
"step": 98000,
|
| 1079 |
+
"training_step_in_ms": 965.2209133654833
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"_prepare_inputs_in_ms": 4.132391892373562,
|
| 1083 |
+
"compute_loss_in_ms": 268.32156636565924,
|
| 1084 |
+
"epoch": 40.93,
|
| 1085 |
+
"learning_rate/full": 0.00020377031677881017,
|
| 1086 |
+
"loss": 3.6386,
|
| 1087 |
+
"step": 99000,
|
| 1088 |
+
"training_step_in_ms": 965.1960897520185
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"_prepare_inputs_in_ms": 4.139371138066053,
|
| 1092 |
+
"compute_loss_in_ms": 268.1989936903119,
|
| 1093 |
+
"epoch": 41.34,
|
| 1094 |
+
"learning_rate/full": 0.00020062894644154732,
|
| 1095 |
+
"loss": 3.628,
|
| 1096 |
+
"step": 100000,
|
| 1097 |
+
"training_step_in_ms": 969.4596163183451
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 41.34,
|
| 1101 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.565774440765381,
|
| 1102 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.26571601552352997,
|
| 1103 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.6144,
|
| 1104 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.307,
|
| 1105 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.322,
|
| 1106 |
+
"step": 100000
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"_prepare_inputs_in_ms": 4.108485870245026,
|
| 1110 |
+
"compute_loss_in_ms": 267.23163178935647,
|
| 1111 |
+
"epoch": 41.75,
|
| 1112 |
+
"learning_rate/full": 0.00019748742092116103,
|
| 1113 |
+
"loss": 3.6364,
|
| 1114 |
+
"step": 101000,
|
| 1115 |
+
"training_step_in_ms": 962.6803079359233
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"_prepare_inputs_in_ms": 4.145125687122345,
|
| 1119 |
+
"compute_loss_in_ms": 267.68679490312934,
|
| 1120 |
+
"epoch": 42.17,
|
| 1121 |
+
"learning_rate/full": 0.00019434651534206603,
|
| 1122 |
+
"loss": 3.6315,
|
| 1123 |
+
"step": 102000,
|
| 1124 |
+
"training_step_in_ms": 964.9059623852372
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"_prepare_inputs_in_ms": 4.14092817902565,
|
| 1128 |
+
"compute_loss_in_ms": 268.40766886249185,
|
| 1129 |
+
"epoch": 42.58,
|
| 1130 |
+
"learning_rate/full": 0.00019120700467571537,
|
| 1131 |
+
"loss": 3.6249,
|
| 1132 |
+
"step": 103000,
|
| 1133 |
+
"training_step_in_ms": 968.5381288193166
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"_prepare_inputs_in_ms": 4.131742633879185,
|
| 1137 |
+
"compute_loss_in_ms": 267.6234121248126,
|
| 1138 |
+
"epoch": 42.99,
|
| 1139 |
+
"learning_rate/full": 0.00018806966354938863,
|
| 1140 |
+
"loss": 3.6302,
|
| 1141 |
+
"step": 104000,
|
| 1142 |
+
"training_step_in_ms": 965.0357882864773
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"_prepare_inputs_in_ms": 4.159636992961168,
|
| 1146 |
+
"compute_loss_in_ms": 268.2043272703886,
|
| 1147 |
+
"epoch": 43.41,
|
| 1148 |
+
"learning_rate/full": 0.0001849352660550636,
|
| 1149 |
+
"loss": 3.6221,
|
| 1150 |
+
"step": 105000,
|
| 1151 |
+
"training_step_in_ms": 966.8701088428497
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 43.41,
|
| 1155 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.560976266860962,
|
| 1156 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2679473083105533,
|
| 1157 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.6952,
|
| 1158 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.166,
|
| 1159 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.318,
|
| 1160 |
+
"step": 105000
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"_prepare_inputs_in_ms": 4.1656713769203275,
|
| 1164 |
+
"compute_loss_in_ms": 267.7097022458911,
|
| 1165 |
+
"epoch": 43.82,
|
| 1166 |
+
"learning_rate/full": 0.00018180458555842107,
|
| 1167 |
+
"loss": 3.6281,
|
| 1168 |
+
"step": 106000,
|
| 1169 |
+
"training_step_in_ms": 960.9392982535064
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"_prepare_inputs_in_ms": 4.1548894718289375,
|
| 1173 |
+
"compute_loss_in_ms": 268.3351138718426,
|
| 1174 |
+
"epoch": 44.23,
|
| 1175 |
+
"learning_rate/full": 0.00017867839450802815,
|
| 1176 |
+
"loss": 3.6257,
|
| 1177 |
+
"step": 107000,
|
| 1178 |
+
"training_step_in_ms": 967.5997758358717
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"_prepare_inputs_in_ms": 4.136493876576424,
|
| 1182 |
+
"compute_loss_in_ms": 267.79356829449534,
|
| 1183 |
+
"epoch": 44.65,
|
| 1184 |
+
"learning_rate/full": 0.0001755574642447484,
|
| 1185 |
+
"loss": 3.6275,
|
| 1186 |
+
"step": 108000,
|
| 1187 |
+
"training_step_in_ms": 964.6688169278204
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"_prepare_inputs_in_ms": 4.146924342960119,
|
| 1191 |
+
"compute_loss_in_ms": 268.7044747136533,
|
| 1192 |
+
"epoch": 45.06,
|
| 1193 |
+
"learning_rate/full": 0.00017244256481142465,
|
| 1194 |
+
"loss": 3.6272,
|
| 1195 |
+
"step": 109000,
|
| 1196 |
+
"training_step_in_ms": 966.0691562928259
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"_prepare_inputs_in_ms": 4.129870388656855,
|
| 1200 |
+
"compute_loss_in_ms": 267.92700193077326,
|
| 1201 |
+
"epoch": 45.47,
|
| 1202 |
+
"learning_rate/full": 0.00016933446476288295,
|
| 1203 |
+
"loss": 3.6231,
|
| 1204 |
+
"step": 110000,
|
| 1205 |
+
"training_step_in_ms": 968.8978024721146
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 45.47,
|
| 1209 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.557985544204712,
|
| 1210 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.26707363597972006,
|
| 1211 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.4146,
|
| 1212 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.334,
|
| 1213 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.323,
|
| 1214 |
+
"step": 110000
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"_prepare_inputs_in_ms": 4.14337538001014,
|
| 1218 |
+
"compute_loss_in_ms": 267.68642891570926,
|
| 1219 |
+
"epoch": 45.89,
|
| 1220 |
+
"learning_rate/full": 0.0001662308313853947,
|
| 1221 |
+
"loss": 3.6179,
|
| 1222 |
+
"step": 111000,
|
| 1223 |
+
"training_step_in_ms": 963.7148243077099
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"_prepare_inputs_in_ms": 4.14999657869339,
|
| 1227 |
+
"compute_loss_in_ms": 267.70868534594774,
|
| 1228 |
+
"epoch": 46.3,
|
| 1229 |
+
"learning_rate/full": 0.000163135546733723,
|
| 1230 |
+
"loss": 3.6124,
|
| 1231 |
+
"step": 112000,
|
| 1232 |
+
"training_step_in_ms": 967.6017691344023
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"_prepare_inputs_in_ms": 4.146619476377964,
|
| 1236 |
+
"compute_loss_in_ms": 268.191356562078,
|
| 1237 |
+
"epoch": 46.71,
|
| 1238 |
+
"learning_rate/full": 0.00016005245741683915,
|
| 1239 |
+
"loss": 3.62,
|
| 1240 |
+
"step": 113000,
|
| 1241 |
+
"training_step_in_ms": 966.300628580153
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"_prepare_inputs_in_ms": 4.153369773179293,
|
| 1245 |
+
"compute_loss_in_ms": 267.8536421582103,
|
| 1246 |
+
"epoch": 47.13,
|
| 1247 |
+
"learning_rate/full": 0.00015697922455833988,
|
| 1248 |
+
"loss": 3.6157,
|
| 1249 |
+
"step": 114000,
|
| 1250 |
+
"training_step_in_ms": 964.6176136285067
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"_prepare_inputs_in_ms": 4.187301464378834,
|
| 1254 |
+
"compute_loss_in_ms": 268.31423101201653,
|
| 1255 |
+
"epoch": 47.54,
|
| 1256 |
+
"learning_rate/full": 0.0001539166064324471,
|
| 1257 |
+
"loss": 3.6143,
|
| 1258 |
+
"step": 115000,
|
| 1259 |
+
"training_step_in_ms": 968.9795580692589
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 47.54,
|
| 1263 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.556086778640747,
|
| 1264 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2680325079129448,
|
| 1265 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.6426,
|
| 1266 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.173,
|
| 1267 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.318,
|
| 1268 |
+
"step": 115000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"_prepare_inputs_in_ms": 4.146232630421475,
|
| 1272 |
+
"compute_loss_in_ms": 267.5073589235544,
|
| 1273 |
+
"epoch": 47.95,
|
| 1274 |
+
"learning_rate/full": 0.00015086535869435647,
|
| 1275 |
+
"loss": 3.6143,
|
| 1276 |
+
"step": 116000,
|
| 1277 |
+
"training_step_in_ms": 961.0740608982742
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"_prepare_inputs_in_ms": 4.149144068360329,
|
| 1281 |
+
"compute_loss_in_ms": 268.67081797868013,
|
| 1282 |
+
"epoch": 48.37,
|
| 1283 |
+
"learning_rate/full": 0.00014782623419379065,
|
| 1284 |
+
"loss": 3.6076,
|
| 1285 |
+
"step": 117000,
|
| 1286 |
+
"training_step_in_ms": 968.1870553046465
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"_prepare_inputs_in_ms": 4.165053818374872,
|
| 1290 |
+
"compute_loss_in_ms": 267.5217378772795,
|
| 1291 |
+
"epoch": 48.78,
|
| 1292 |
+
"learning_rate/full": 0.00014479998278924466,
|
| 1293 |
+
"loss": 3.6033,
|
| 1294 |
+
"step": 118000,
|
| 1295 |
+
"training_step_in_ms": 964.0207477062941
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"_prepare_inputs_in_ms": 4.148088995367289,
|
| 1299 |
+
"compute_loss_in_ms": 267.9359416142106,
|
| 1300 |
+
"epoch": 49.19,
|
| 1301 |
+
"learning_rate/full": 0.00014178735116296984,
|
| 1302 |
+
"loss": 3.6099,
|
| 1303 |
+
"step": 119000,
|
| 1304 |
+
"training_step_in_ms": 965.35854877159
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"_prepare_inputs_in_ms": 4.147441305220127,
|
| 1308 |
+
"compute_loss_in_ms": 268.18652522563934,
|
| 1309 |
+
"epoch": 49.61,
|
| 1310 |
+
"learning_rate/full": 0.00013878908263674099,
|
| 1311 |
+
"loss": 3.6041,
|
| 1312 |
+
"step": 120000,
|
| 1313 |
+
"training_step_in_ms": 966.7598981000483
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 49.61,
|
| 1317 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5552496910095215,
|
| 1318 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2690825925881079,
|
| 1319 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.206,
|
| 1320 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.362,
|
| 1321 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.324,
|
| 1322 |
+
"step": 120000
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"_prepare_inputs_in_ms": 4.151007533073425,
|
| 1326 |
+
"compute_loss_in_ms": 267.91314566135406,
|
| 1327 |
+
"epoch": 50.02,
|
| 1328 |
+
"learning_rate/full": 0.00013580293864950247,
|
| 1329 |
+
"loss": 3.6068,
|
| 1330 |
+
"step": 121000,
|
| 1331 |
+
"training_step_in_ms": 965.1756884045899
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"_prepare_inputs_in_ms": 4.154482748359442,
|
| 1335 |
+
"compute_loss_in_ms": 268.1262241154909,
|
| 1336 |
+
"epoch": 50.43,
|
| 1337 |
+
"learning_rate/full": 0.00013283562815289706,
|
| 1338 |
+
"loss": 3.5972,
|
| 1339 |
+
"step": 122000,
|
| 1340 |
+
"training_step_in_ms": 965.2079959511757
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"_prepare_inputs_in_ms": 4.1559099070727825,
|
| 1344 |
+
"compute_loss_in_ms": 267.7796282917261,
|
| 1345 |
+
"epoch": 50.85,
|
| 1346 |
+
"learning_rate/full": 0.0001298848894600429,
|
| 1347 |
+
"loss": 3.6031,
|
| 1348 |
+
"step": 123000,
|
| 1349 |
+
"training_step_in_ms": 966.0297281630337
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"_prepare_inputs_in_ms": 4.1683206632733345,
|
| 1353 |
+
"compute_loss_in_ms": 268.80906841158867,
|
| 1354 |
+
"epoch": 51.26,
|
| 1355 |
+
"learning_rate/full": 0.00012694852315582903,
|
| 1356 |
+
"loss": 3.5985,
|
| 1357 |
+
"step": 124000,
|
| 1358 |
+
"training_step_in_ms": 967.5091603025794
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"_prepare_inputs_in_ms": 4.137588050216436,
|
| 1362 |
+
"compute_loss_in_ms": 267.6871258877218,
|
| 1363 |
+
"epoch": 51.67,
|
| 1364 |
+
"learning_rate/full": 0.00012403603541958643,
|
| 1365 |
+
"loss": 3.6029,
|
| 1366 |
+
"step": 125000,
|
| 1367 |
+
"training_step_in_ms": 964.4286920540035
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 51.67,
|
| 1371 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5499510765075684,
|
| 1372 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.268844144332836,
|
| 1373 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.7216,
|
| 1374 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.293,
|
| 1375 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.322,
|
| 1376 |
+
"step": 125000
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"_prepare_inputs_in_ms": 4.127907207826289,
|
| 1380 |
+
"compute_loss_in_ms": 267.8013560883701,
|
| 1381 |
+
"epoch": 52.09,
|
| 1382 |
+
"learning_rate/full": 0.00012113936318920029,
|
| 1383 |
+
"loss": 3.5962,
|
| 1384 |
+
"step": 126000,
|
| 1385 |
+
"training_step_in_ms": 962.9286280833185
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"_prepare_inputs_in_ms": 4.151306886225939,
|
| 1389 |
+
"compute_loss_in_ms": 267.83537547290325,
|
| 1390 |
+
"epoch": 52.5,
|
| 1391 |
+
"learning_rate/full": 0.00011825927853225391,
|
| 1392 |
+
"loss": 3.5974,
|
| 1393 |
+
"step": 127000,
|
| 1394 |
+
"training_step_in_ms": 967.7427954226732
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"_prepare_inputs_in_ms": 4.118186391890049,
|
| 1398 |
+
"compute_loss_in_ms": 268.2419737614691,
|
| 1399 |
+
"epoch": 52.91,
|
| 1400 |
+
"learning_rate/full": 0.00011540225211753402,
|
| 1401 |
+
"loss": 3.5979,
|
| 1402 |
+
"step": 128000,
|
| 1403 |
+
"training_step_in_ms": 963.5640154518187
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"_prepare_inputs_in_ms": 4.146437399089336,
|
| 1407 |
+
"compute_loss_in_ms": 267.88536206260324,
|
| 1408 |
+
"epoch": 53.33,
|
| 1409 |
+
"learning_rate/full": 0.00011256609893124084,
|
| 1410 |
+
"loss": 3.5919,
|
| 1411 |
+
"step": 129000,
|
| 1412 |
+
"training_step_in_ms": 965.4581209644675
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"_prepare_inputs_in_ms": 4.140674088150263,
|
| 1416 |
+
"compute_loss_in_ms": 267.7817959152162,
|
| 1417 |
+
"epoch": 53.74,
|
| 1418 |
+
"learning_rate/full": 0.00010975151875173475,
|
| 1419 |
+
"loss": 3.5954,
|
| 1420 |
+
"step": 130000,
|
| 1421 |
+
"training_step_in_ms": 967.2819016650319
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"epoch": 53.74,
|
| 1425 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5497984886169434,
|
| 1426 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.27146193885645503,
|
| 1427 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.8468,
|
| 1428 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.277,
|
| 1429 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.321,
|
| 1430 |
+
"step": 130000
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"_prepare_inputs_in_ms": 4.144404723149974,
|
| 1434 |
+
"compute_loss_in_ms": 267.8810519017279,
|
| 1435 |
+
"epoch": 54.15,
|
| 1436 |
+
"learning_rate/full": 0.00010695920603455975,
|
| 1437 |
+
"loss": 3.5914,
|
| 1438 |
+
"step": 131000,
|
| 1439 |
+
"training_step_in_ms": 962.7643201723695
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"_prepare_inputs_in_ms": 4.15174587816,
|
| 1443 |
+
"compute_loss_in_ms": 268.55379743501544,
|
| 1444 |
+
"epoch": 54.57,
|
| 1445 |
+
"learning_rate/full": 0.00010418984974109642,
|
| 1446 |
+
"loss": 3.5916,
|
| 1447 |
+
"step": 132000,
|
| 1448 |
+
"training_step_in_ms": 965.8136657737195
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"_prepare_inputs_in_ms": 4.164624106138945,
|
| 1452 |
+
"compute_loss_in_ms": 268.1068575233221,
|
| 1453 |
+
"epoch": 54.98,
|
| 1454 |
+
"learning_rate/full": 0.00010144413316857143,
|
| 1455 |
+
"loss": 3.5824,
|
| 1456 |
+
"step": 133000,
|
| 1457 |
+
"training_step_in_ms": 961.5831676833332
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"_prepare_inputs_in_ms": 4.1917338743805885,
|
| 1461 |
+
"compute_loss_in_ms": 268.29790291562676,
|
| 1462 |
+
"epoch": 55.39,
|
| 1463 |
+
"learning_rate/full": 9.872273378146393e-05,
|
| 1464 |
+
"loss": 3.5822,
|
| 1465 |
+
"step": 134000,
|
| 1466 |
+
"training_step_in_ms": 967.4604325480759
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"_prepare_inputs_in_ms": 4.154745582491159,
|
| 1470 |
+
"compute_loss_in_ms": 268.35503727942705,
|
| 1471 |
+
"epoch": 55.81,
|
| 1472 |
+
"learning_rate/full": 9.602632304435166e-05,
|
| 1473 |
+
"loss": 3.5837,
|
| 1474 |
+
"step": 135000,
|
| 1475 |
+
"training_step_in_ms": 963.2355434708297
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 55.81,
|
| 1479 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5466184616088867,
|
| 1480 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2701472521164958,
|
| 1481 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.7707,
|
| 1482 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.287,
|
| 1483 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.321,
|
| 1484 |
+
"step": 135000
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"_prepare_inputs_in_ms": 4.1385579000159005,
|
| 1488 |
+
"compute_loss_in_ms": 268.1807456240058,
|
| 1489 |
+
"epoch": 56.22,
|
| 1490 |
+
"learning_rate/full": 9.335556625623667e-05,
|
| 1491 |
+
"loss": 3.5886,
|
| 1492 |
+
"step": 136000,
|
| 1493 |
+
"training_step_in_ms": 964.7279985249043
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"_prepare_inputs_in_ms": 4.140180014073849,
|
| 1497 |
+
"compute_loss_in_ms": 268.0731739960611,
|
| 1498 |
+
"epoch": 56.63,
|
| 1499 |
+
"learning_rate/full": 9.071112238639306e-05,
|
| 1500 |
+
"loss": 3.5841,
|
| 1501 |
+
"step": 137000,
|
| 1502 |
+
"training_step_in_ms": 964.2245756573975
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"_prepare_inputs_in_ms": 4.159709714353085,
|
| 1506 |
+
"compute_loss_in_ms": 268.24955869838595,
|
| 1507 |
+
"epoch": 57.05,
|
| 1508 |
+
"learning_rate/full": 8.809103753928327e-05,
|
| 1509 |
+
"loss": 3.5843,
|
| 1510 |
+
"step": 138000,
|
| 1511 |
+
"training_step_in_ms": 966.5491472817957
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"_prepare_inputs_in_ms": 4.165288005024195,
|
| 1515 |
+
"compute_loss_in_ms": 268.81143694743514,
|
| 1516 |
+
"epoch": 57.46,
|
| 1517 |
+
"learning_rate/full": 8.550119824369325e-05,
|
| 1518 |
+
"loss": 3.5867,
|
| 1519 |
+
"step": 139000,
|
| 1520 |
+
"training_step_in_ms": 964.0126786530018
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"_prepare_inputs_in_ms": 4.144395582377911,
|
| 1524 |
+
"compute_loss_in_ms": 268.206242531538,
|
| 1525 |
+
"epoch": 57.88,
|
| 1526 |
+
"learning_rate/full": 8.29370600270935e-05,
|
| 1527 |
+
"loss": 3.5852,
|
| 1528 |
+
"step": 140000,
|
| 1529 |
+
"training_step_in_ms": 962.685001052916
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 57.88,
|
| 1533 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5455029010772705,
|
| 1534 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2711795494440531,
|
| 1535 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.9766,
|
| 1536 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.259,
|
| 1537 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.321,
|
| 1538 |
+
"step": 140000
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"_prepare_inputs_in_ms": 4.152141958475113,
|
| 1542 |
+
"compute_loss_in_ms": 267.9933222196996,
|
| 1543 |
+
"epoch": 58.29,
|
| 1544 |
+
"learning_rate/full": 8.040438372331344e-05,
|
| 1545 |
+
"loss": 3.5737,
|
| 1546 |
+
"step": 141000,
|
| 1547 |
+
"training_step_in_ms": 965.1287141442299
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"_prepare_inputs_in_ms": 4.142403397709131,
|
| 1551 |
+
"compute_loss_in_ms": 267.5065658353269,
|
| 1552 |
+
"epoch": 58.7,
|
| 1553 |
+
"learning_rate/full": 7.790121584830201e-05,
|
| 1554 |
+
"loss": 3.575,
|
| 1555 |
+
"step": 142000,
|
| 1556 |
+
"training_step_in_ms": 962.7664158046246
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"_prepare_inputs_in_ms": 4.150233589112759,
|
| 1560 |
+
"compute_loss_in_ms": 268.65408623218536,
|
| 1561 |
+
"epoch": 59.12,
|
| 1562 |
+
"learning_rate/full": 7.542817402127658e-05,
|
| 1563 |
+
"loss": 3.5709,
|
| 1564 |
+
"step": 143000,
|
| 1565 |
+
"training_step_in_ms": 964.2099178209901
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"_prepare_inputs_in_ms": 4.130799826234579,
|
| 1569 |
+
"compute_loss_in_ms": 267.8640896603465,
|
| 1570 |
+
"epoch": 59.53,
|
| 1571 |
+
"learning_rate/full": 7.298586842830323e-05,
|
| 1572 |
+
"loss": 3.5775,
|
| 1573 |
+
"step": 144000,
|
| 1574 |
+
"training_step_in_ms": 964.5784216374159
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"_prepare_inputs_in_ms": 4.146471511572599,
|
| 1578 |
+
"compute_loss_in_ms": 268.14434216171503,
|
| 1579 |
+
"epoch": 59.94,
|
| 1580 |
+
"learning_rate/full": 7.057490167174197e-05,
|
| 1581 |
+
"loss": 3.5781,
|
| 1582 |
+
"step": 145000,
|
| 1583 |
+
"training_step_in_ms": 964.2295859828591
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 59.94,
|
| 1587 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5419652462005615,
|
| 1588 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2710637844956296,
|
| 1589 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.2475,
|
| 1590 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.224,
|
| 1591 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.319,
|
| 1592 |
+
"step": 145000
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"_prepare_inputs_in_ms": 4.13495619122575,
|
| 1596 |
+
"compute_loss_in_ms": 267.6772438697517,
|
| 1597 |
+
"epoch": 60.36,
|
| 1598 |
+
"learning_rate/full": 6.819586862156388e-05,
|
| 1599 |
+
"loss": 3.5713,
|
| 1600 |
+
"step": 146000,
|
| 1601 |
+
"training_step_in_ms": 963.0674764961004
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"_prepare_inputs_in_ms": 4.138938769698143,
|
| 1605 |
+
"compute_loss_in_ms": 268.4652929417789,
|
| 1606 |
+
"epoch": 60.77,
|
| 1607 |
+
"learning_rate/full": 6.58493562685758e-05,
|
| 1608 |
+
"loss": 3.5755,
|
| 1609 |
+
"step": 147000,
|
| 1610 |
+
"training_step_in_ms": 965.1382315270603
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"_prepare_inputs_in_ms": 4.149636901915073,
|
| 1614 |
+
"compute_loss_in_ms": 268.6144716888666,
|
| 1615 |
+
"epoch": 61.18,
|
| 1616 |
+
"learning_rate/full": 6.35336446255852e-05,
|
| 1617 |
+
"loss": 3.5689,
|
| 1618 |
+
"step": 148000,
|
| 1619 |
+
"training_step_in_ms": 967.6213804855943
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"_prepare_inputs_in_ms": 4.154591448605061,
|
| 1623 |
+
"compute_loss_in_ms": 268.10440719127655,
|
| 1624 |
+
"epoch": 61.6,
|
| 1625 |
+
"learning_rate/full": 6.125393638794017e-05,
|
| 1626 |
+
"loss": 3.5736,
|
| 1627 |
+
"step": 149000,
|
| 1628 |
+
"training_step_in_ms": 966.7320594601333
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"_prepare_inputs_in_ms": 4.149040505290031,
|
| 1632 |
+
"compute_loss_in_ms": 268.75643199309707,
|
| 1633 |
+
"epoch": 62.01,
|
| 1634 |
+
"learning_rate/full": 5.900623127984053e-05,
|
| 1635 |
+
"loss": 3.57,
|
| 1636 |
+
"step": 150000,
|
| 1637 |
+
"training_step_in_ms": 965.394243825227
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 62.01,
|
| 1641 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.542189359664917,
|
| 1642 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.27182213175584513,
|
| 1643 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.2758,
|
| 1644 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.22,
|
| 1645 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.319,
|
| 1646 |
+
"step": 150000
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"_prepare_inputs_in_ms": 4.197841770038372,
|
| 1650 |
+
"compute_loss_in_ms": 268.5856811106205,
|
| 1651 |
+
"epoch": 62.42,
|
| 1652 |
+
"learning_rate/full": 5.6795579206763614e-05,
|
| 1653 |
+
"loss": 3.5655,
|
| 1654 |
+
"step": 151000,
|
| 1655 |
+
"training_step_in_ms": 963.427967004478
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"_prepare_inputs_in_ms": 4.1404072009027,
|
| 1659 |
+
"compute_loss_in_ms": 267.94744442403316,
|
| 1660 |
+
"epoch": 62.84,
|
| 1661 |
+
"learning_rate/full": 5.462026068170363e-05,
|
| 1662 |
+
"loss": 3.5665,
|
| 1663 |
+
"step": 152000,
|
| 1664 |
+
"training_step_in_ms": 965.5440159775317
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"_prepare_inputs_in_ms": 4.148986879736185,
|
| 1668 |
+
"compute_loss_in_ms": 268.1357101947069,
|
| 1669 |
+
"epoch": 63.25,
|
| 1670 |
+
"learning_rate/full": 5.247868899032384e-05,
|
| 1671 |
+
"loss": 3.5648,
|
| 1672 |
+
"step": 153000,
|
| 1673 |
+
"training_step_in_ms": 965.260343439877
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"_prepare_inputs_in_ms": 4.144355583935976,
|
| 1677 |
+
"compute_loss_in_ms": 268.8695700503886,
|
| 1678 |
+
"epoch": 63.66,
|
| 1679 |
+
"learning_rate/full": 5.0375675588795876e-05,
|
| 1680 |
+
"loss": 3.5699,
|
| 1681 |
+
"step": 154000,
|
| 1682 |
+
"training_step_in_ms": 968.1045257672668
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"_prepare_inputs_in_ms": 4.134350396692753,
|
| 1686 |
+
"compute_loss_in_ms": 267.78631913661957,
|
| 1687 |
+
"epoch": 64.08,
|
| 1688 |
+
"learning_rate/full": 4.830957975043959e-05,
|
| 1689 |
+
"loss": 3.5654,
|
| 1690 |
+
"step": 155000,
|
| 1691 |
+
"training_step_in_ms": 965.367557708174
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 64.08,
|
| 1695 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5409913063049316,
|
| 1696 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.273054083346476,
|
| 1697 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.4578,
|
| 1698 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.197,
|
| 1699 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.319,
|
| 1700 |
+
"step": 155000
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"_prepare_inputs_in_ms": 4.169517766411712,
|
| 1704 |
+
"compute_loss_in_ms": 267.94907980412245,
|
| 1705 |
+
"epoch": 64.49,
|
| 1706 |
+
"learning_rate/full": 4.628091125348743e-05,
|
| 1707 |
+
"loss": 3.562,
|
| 1708 |
+
"step": 156000,
|
| 1709 |
+
"training_step_in_ms": 964.7187770940363
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"_prepare_inputs_in_ms": 4.127725187689066,
|
| 1713 |
+
"compute_loss_in_ms": 267.56007508188486,
|
| 1714 |
+
"epoch": 64.9,
|
| 1715 |
+
"learning_rate/full": 4.429017064153536e-05,
|
| 1716 |
+
"loss": 3.5599,
|
| 1717 |
+
"step": 157000,
|
| 1718 |
+
"training_step_in_ms": 961.402901135385
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"_prepare_inputs_in_ms": 4.122306831181049,
|
| 1722 |
+
"compute_loss_in_ms": 268.3586079515517,
|
| 1723 |
+
"epoch": 65.32,
|
| 1724 |
+
"learning_rate/full": 4.233784910004124e-05,
|
| 1725 |
+
"loss": 3.5666,
|
| 1726 |
+
"step": 158000,
|
| 1727 |
+
"training_step_in_ms": 965.8669985719025
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"_prepare_inputs_in_ms": 4.136414989829063,
|
| 1731 |
+
"compute_loss_in_ms": 267.9736096225679,
|
| 1732 |
+
"epoch": 65.73,
|
| 1733 |
+
"learning_rate/full": 4.0424428335132335e-05,
|
| 1734 |
+
"loss": 3.5573,
|
| 1735 |
+
"step": 159000,
|
| 1736 |
+
"training_step_in_ms": 965.0534134693444
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"_prepare_inputs_in_ms": 4.14548010751605,
|
| 1740 |
+
"compute_loss_in_ms": 268.10164315626025,
|
| 1741 |
+
"epoch": 66.14,
|
| 1742 |
+
"learning_rate/full": 3.855038045475119e-05,
|
| 1743 |
+
"loss": 3.5569,
|
| 1744 |
+
"step": 160000,
|
| 1745 |
+
"training_step_in_ms": 965.0257755257189
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 66.14,
|
| 1749 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.540762424468994,
|
| 1750 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2735980306318844,
|
| 1751 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.6347,
|
| 1752 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.305,
|
| 1753 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.322,
|
| 1754 |
+
"step": 160000
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"_prepare_inputs_in_ms": 4.134731147347427,
|
| 1758 |
+
"compute_loss_in_ms": 268.0676885545254,
|
| 1759 |
+
"epoch": 66.56,
|
| 1760 |
+
"learning_rate/full": 3.671616785217033e-05,
|
| 1761 |
+
"loss": 3.559,
|
| 1762 |
+
"step": 161000,
|
| 1763 |
+
"training_step_in_ms": 961.5968884006143
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"_prepare_inputs_in_ms": 4.13828482478857,
|
| 1767 |
+
"compute_loss_in_ms": 268.11520731821656,
|
| 1768 |
+
"epoch": 66.97,
|
| 1769 |
+
"learning_rate/full": 3.4920467704438286e-05,
|
| 1770 |
+
"loss": 3.5617,
|
| 1771 |
+
"step": 162000,
|
| 1772 |
+
"training_step_in_ms": 963.3356633149087
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"_prepare_inputs_in_ms": 4.1371137127280235,
|
| 1776 |
+
"compute_loss_in_ms": 267.97775723040104,
|
| 1777 |
+
"epoch": 67.38,
|
| 1778 |
+
"learning_rate/full": 3.3169048798042254e-05,
|
| 1779 |
+
"loss": 3.5629,
|
| 1780 |
+
"step": 163000,
|
| 1781 |
+
"training_step_in_ms": 964.7222346775234
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"_prepare_inputs_in_ms": 4.129249203950167,
|
| 1785 |
+
"compute_loss_in_ms": 267.63603001460433,
|
| 1786 |
+
"epoch": 67.8,
|
| 1787 |
+
"learning_rate/full": 3.145532456480391e-05,
|
| 1788 |
+
"loss": 3.5596,
|
| 1789 |
+
"step": 164000,
|
| 1790 |
+
"training_step_in_ms": 964.5063005648553
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"_prepare_inputs_in_ms": 4.151564922183752,
|
| 1794 |
+
"compute_loss_in_ms": 268.839259788394,
|
| 1795 |
+
"epoch": 68.21,
|
| 1796 |
+
"learning_rate/full": 2.9784920606062528e-05,
|
| 1797 |
+
"loss": 3.5602,
|
| 1798 |
+
"step": 165000,
|
| 1799 |
+
"training_step_in_ms": 972.3141440451145
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 68.21,
|
| 1803 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.539008855819702,
|
| 1804 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2725576622048259,
|
| 1805 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.4123,
|
| 1806 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.202,
|
| 1807 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.319,
|
| 1808 |
+
"step": 165000
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"_prepare_inputs_in_ms": 4.126259837208725,
|
| 1812 |
+
"compute_loss_in_ms": 267.8272004313767,
|
| 1813 |
+
"epoch": 68.62,
|
| 1814 |
+
"learning_rate/full": 2.8156514671178745e-05,
|
| 1815 |
+
"loss": 3.5603,
|
| 1816 |
+
"step": 166000,
|
| 1817 |
+
"training_step_in_ms": 964.2809295020998
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"_prepare_inputs_in_ms": 4.134287599474192,
|
| 1821 |
+
"compute_loss_in_ms": 267.94721764326096,
|
| 1822 |
+
"epoch": 69.04,
|
| 1823 |
+
"learning_rate/full": 2.65705085449506e-05,
|
| 1824 |
+
"loss": 3.5591,
|
| 1825 |
+
"step": 167000,
|
| 1826 |
+
"training_step_in_ms": 963.7021813839674
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"_prepare_inputs_in_ms": 4.147611241787672,
|
| 1830 |
+
"compute_loss_in_ms": 268.29229406639934,
|
| 1831 |
+
"epoch": 69.45,
|
| 1832 |
+
"learning_rate/full": 2.5025770357450595e-05,
|
| 1833 |
+
"loss": 3.5561,
|
| 1834 |
+
"step": 168000,
|
| 1835 |
+
"training_step_in_ms": 967.2244190610945
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"_prepare_inputs_in_ms": 4.15412675216794,
|
| 1839 |
+
"compute_loss_in_ms": 268.25271063297987,
|
| 1840 |
+
"epoch": 69.86,
|
| 1841 |
+
"learning_rate/full": 2.352577066262569e-05,
|
| 1842 |
+
"loss": 3.5576,
|
| 1843 |
+
"step": 169000,
|
| 1844 |
+
"training_step_in_ms": 965.3532739318907
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"_prepare_inputs_in_ms": 4.129838448017836,
|
| 1848 |
+
"compute_loss_in_ms": 267.92896181344986,
|
| 1849 |
+
"epoch": 70.28,
|
| 1850 |
+
"learning_rate/full": 2.206931334324922e-05,
|
| 1851 |
+
"loss": 3.5536,
|
| 1852 |
+
"step": 170000,
|
| 1853 |
+
"training_step_in_ms": 965.5082765445113
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 70.28,
|
| 1857 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5376806259155273,
|
| 1858 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2738309179784362,
|
| 1859 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 77.9139,
|
| 1860 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.268,
|
| 1861 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.321,
|
| 1862 |
+
"step": 170000
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"_prepare_inputs_in_ms": 4.163786387298165,
|
| 1866 |
+
"compute_loss_in_ms": 267.95297726243734,
|
| 1867 |
+
"epoch": 70.69,
|
| 1868 |
+
"learning_rate/full": 2.0656757758371282e-05,
|
| 1869 |
+
"loss": 3.5571,
|
| 1870 |
+
"step": 171000,
|
| 1871 |
+
"training_step_in_ms": 960.7510039620101
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"_prepare_inputs_in_ms": 4.137572426348925,
|
| 1875 |
+
"compute_loss_in_ms": 267.8121683038771,
|
| 1876 |
+
"epoch": 71.1,
|
| 1877 |
+
"learning_rate/full": 1.9287105043128472e-05,
|
| 1878 |
+
"loss": 3.555,
|
| 1879 |
+
"step": 172000,
|
| 1880 |
+
"training_step_in_ms": 963.8648240976036
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"_prepare_inputs_in_ms": 4.139789171516895,
|
| 1884 |
+
"compute_loss_in_ms": 268.3458735384047,
|
| 1885 |
+
"epoch": 71.52,
|
| 1886 |
+
"learning_rate/full": 1.796343238799574e-05,
|
| 1887 |
+
"loss": 3.5518,
|
| 1888 |
+
"step": 173000,
|
| 1889 |
+
"training_step_in_ms": 965.790959071368
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"_prepare_inputs_in_ms": 4.144272416830063,
|
| 1893 |
+
"compute_loss_in_ms": 268.10323084518313,
|
| 1894 |
+
"epoch": 71.93,
|
| 1895 |
+
"learning_rate/full": 1.6684674532049582e-05,
|
| 1896 |
+
"loss": 3.5512,
|
| 1897 |
+
"step": 174000,
|
| 1898 |
+
"training_step_in_ms": 961.3717007525265
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"_prepare_inputs_in_ms": 4.1666854321956635,
|
| 1902 |
+
"compute_loss_in_ms": 268.03433157876134,
|
| 1903 |
+
"epoch": 72.34,
|
| 1904 |
+
"learning_rate/full": 1.5451146989656617e-05,
|
| 1905 |
+
"loss": 3.5495,
|
| 1906 |
+
"step": 175000,
|
| 1907 |
+
"training_step_in_ms": 967.3243609592319
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 72.34,
|
| 1911 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.536776065826416,
|
| 1912 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2743759293675203,
|
| 1913 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.6158,
|
| 1914 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.176,
|
| 1915 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.318,
|
| 1916 |
+
"step": 175000
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"_prepare_inputs_in_ms": 4.134207051701662,
|
| 1920 |
+
"compute_loss_in_ms": 267.52714550867677,
|
| 1921 |
+
"epoch": 72.76,
|
| 1922 |
+
"learning_rate/full": 1.4261987845053304e-05,
|
| 1923 |
+
"loss": 3.5555,
|
| 1924 |
+
"step": 176000,
|
| 1925 |
+
"training_step_in_ms": 962.5389591343701
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"_prepare_inputs_in_ms": 4.146069306880236,
|
| 1929 |
+
"compute_loss_in_ms": 268.7413688749075,
|
| 1930 |
+
"epoch": 73.17,
|
| 1931 |
+
"learning_rate/full": 1.3119868774900613e-05,
|
| 1932 |
+
"loss": 3.5512,
|
| 1933 |
+
"step": 177000,
|
| 1934 |
+
"training_step_in_ms": 964.9299626871943
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"_prepare_inputs_in_ms": 4.148835156112909,
|
| 1938 |
+
"compute_loss_in_ms": 268.0495460778475,
|
| 1939 |
+
"epoch": 73.58,
|
| 1940 |
+
"learning_rate/full": 1.2023859580780273e-05,
|
| 1941 |
+
"loss": 3.5529,
|
| 1942 |
+
"step": 178000,
|
| 1943 |
+
"training_step_in_ms": 967.7268707863986
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"_prepare_inputs_in_ms": 4.1666895635426044,
|
| 1947 |
+
"compute_loss_in_ms": 268.2634797357023,
|
| 1948 |
+
"epoch": 74.0,
|
| 1949 |
+
"learning_rate/full": 1.0973203331088377e-05,
|
| 1950 |
+
"loss": 3.5538,
|
| 1951 |
+
"step": 179000,
|
| 1952 |
+
"training_step_in_ms": 965.1008929647505
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"_prepare_inputs_in_ms": 4.153850518167019,
|
| 1956 |
+
"compute_loss_in_ms": 268.7998457066715,
|
| 1957 |
+
"epoch": 74.41,
|
| 1958 |
+
"learning_rate/full": 9.970260528869224e-06,
|
| 1959 |
+
"loss": 3.5524,
|
| 1960 |
+
"step": 180000,
|
| 1961 |
+
"training_step_in_ms": 968.3517145328224
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 74.41,
|
| 1965 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.536918878555298,
|
| 1966 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2735705193198496,
|
| 1967 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.7713,
|
| 1968 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.156,
|
| 1969 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.317,
|
| 1970 |
+
"step": 180000
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"_prepare_inputs_in_ms": 4.127806843054004,
|
| 1974 |
+
"compute_loss_in_ms": 268.1048993989825,
|
| 1975 |
+
"epoch": 74.82,
|
| 1976 |
+
"learning_rate/full": 9.01420472138852e-06,
|
| 1977 |
+
"loss": 3.5468,
|
| 1978 |
+
"step": 181000,
|
| 1979 |
+
"training_step_in_ms": 960.1139997318387
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"_prepare_inputs_in_ms": 4.13564395532012,
|
| 1983 |
+
"compute_loss_in_ms": 267.3685629181564,
|
| 1984 |
+
"epoch": 75.24,
|
| 1985 |
+
"learning_rate/full": 8.105271801111003e-06,
|
| 1986 |
+
"loss": 3.552,
|
| 1987 |
+
"step": 182000,
|
| 1988 |
+
"training_step_in_ms": 964.2471651136875
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"_prepare_inputs_in_ms": 4.140480011701584,
|
| 1992 |
+
"compute_loss_in_ms": 267.48710445687175,
|
| 1993 |
+
"epoch": 75.65,
|
| 1994 |
+
"learning_rate/full": 7.243686033634145e-06,
|
| 1995 |
+
"loss": 3.5473,
|
| 1996 |
+
"step": 183000,
|
| 1997 |
+
"training_step_in_ms": 966.0506127551198
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"_prepare_inputs_in_ms": 4.12862478941679,
|
| 2001 |
+
"compute_loss_in_ms": 268.38187746331096,
|
| 2002 |
+
"epoch": 76.06,
|
| 2003 |
+
"learning_rate/full": 6.429660002353832e-06,
|
| 2004 |
+
"loss": 3.5535,
|
| 2005 |
+
"step": 184000,
|
| 2006 |
+
"training_step_in_ms": 963.8830341026187
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"_prepare_inputs_in_ms": 4.138918172568083,
|
| 2010 |
+
"compute_loss_in_ms": 267.91103532910347,
|
| 2011 |
+
"epoch": 76.48,
|
| 2012 |
+
"learning_rate/full": 5.663394556012769e-06,
|
| 2013 |
+
"loss": 3.5489,
|
| 2014 |
+
"step": 185000,
|
| 2015 |
+
"training_step_in_ms": 969.0565127506852
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"epoch": 76.48,
|
| 2019 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5358800888061523,
|
| 2020 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2731816800574383,
|
| 2021 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.2007,
|
| 2022 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.23,
|
| 2023 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.32,
|
| 2024 |
+
"step": 185000
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"_prepare_inputs_in_ms": 4.125211863982968,
|
| 2028 |
+
"compute_loss_in_ms": 267.94072189182043,
|
| 2029 |
+
"epoch": 76.89,
|
| 2030 |
+
"learning_rate/full": 4.944383807374142e-06,
|
| 2031 |
+
"loss": 3.5495,
|
| 2032 |
+
"step": 186000,
|
| 2033 |
+
"training_step_in_ms": 963.789376296103
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"_prepare_inputs_in_ms": 4.155690658837557,
|
| 2037 |
+
"compute_loss_in_ms": 267.86934616044164,
|
| 2038 |
+
"epoch": 77.3,
|
| 2039 |
+
"learning_rate/full": 4.273596510048239e-06,
|
| 2040 |
+
"loss": 3.5465,
|
| 2041 |
+
"step": 187000,
|
| 2042 |
+
"training_step_in_ms": 966.8210936710238
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"_prepare_inputs_in_ms": 4.130576055496931,
|
| 2046 |
+
"compute_loss_in_ms": 268.2106507457793,
|
| 2047 |
+
"epoch": 77.72,
|
| 2048 |
+
"learning_rate/full": 3.6517966776118407e-06,
|
| 2049 |
+
"loss": 3.5445,
|
| 2050 |
+
"step": 188000,
|
| 2051 |
+
"training_step_in_ms": 966.5816915780306
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"_prepare_inputs_in_ms": 4.133735220879316,
|
| 2055 |
+
"compute_loss_in_ms": 267.8969533368945,
|
| 2056 |
+
"epoch": 78.13,
|
| 2057 |
+
"learning_rate/full": 3.0789924830065154e-06,
|
| 2058 |
+
"loss": 3.5528,
|
| 2059 |
+
"step": 189000,
|
| 2060 |
+
"training_step_in_ms": 964.6940425820649
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"_prepare_inputs_in_ms": 4.137146957218647,
|
| 2064 |
+
"compute_loss_in_ms": 267.9762873612344,
|
| 2065 |
+
"epoch": 78.54,
|
| 2066 |
+
"learning_rate/full": 2.5536764226682607e-06,
|
| 2067 |
+
"loss": 3.5458,
|
| 2068 |
+
"step": 190000,
|
| 2069 |
+
"training_step_in_ms": 966.734307706356
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 78.54,
|
| 2073 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5369150638580322,
|
| 2074 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.272921503437743,
|
| 2075 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 79.0337,
|
| 2076 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.122,
|
| 2077 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.316,
|
| 2078 |
+
"step": 190000
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"_prepare_inputs_in_ms": 4.113826969774758,
|
| 2082 |
+
"compute_loss_in_ms": 267.3543768711388,
|
| 2083 |
+
"epoch": 78.96,
|
| 2084 |
+
"learning_rate/full": 2.077174866407172e-06,
|
| 2085 |
+
"loss": 3.5455,
|
| 2086 |
+
"step": 191000,
|
| 2087 |
+
"training_step_in_ms": 959.5898663066328
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"_prepare_inputs_in_ms": 4.15868678689003,
|
| 2091 |
+
"compute_loss_in_ms": 268.5514197871089,
|
| 2092 |
+
"epoch": 79.37,
|
| 2093 |
+
"learning_rate/full": 1.650411851111966e-06,
|
| 2094 |
+
"loss": 3.5422,
|
| 2095 |
+
"step": 192000,
|
| 2096 |
+
"training_step_in_ms": 966.848380189389
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"_prepare_inputs_in_ms": 4.152705859392881,
|
| 2100 |
+
"compute_loss_in_ms": 267.7775506339967,
|
| 2101 |
+
"epoch": 79.79,
|
| 2102 |
+
"learning_rate/full": 1.2717824473816864e-06,
|
| 2103 |
+
"loss": 3.5542,
|
| 2104 |
+
"step": 193000,
|
| 2105 |
+
"training_step_in_ms": 966.0438013672829
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"_prepare_inputs_in_ms": 4.149454560130835,
|
| 2109 |
+
"compute_loss_in_ms": 268.18433906137943,
|
| 2110 |
+
"epoch": 80.2,
|
| 2111 |
+
"learning_rate/full": 9.425893981038769e-07,
|
| 2112 |
+
"loss": 3.5513,
|
| 2113 |
+
"step": 194000,
|
| 2114 |
+
"training_step_in_ms": 968.7132156044245
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"_prepare_inputs_in_ms": 4.145272459834814,
|
| 2118 |
+
"compute_loss_in_ms": 269.00753265991807,
|
| 2119 |
+
"epoch": 80.61,
|
| 2120 |
+
"learning_rate/full": 6.625107863321489e-07,
|
| 2121 |
+
"loss": 3.5423,
|
| 2122 |
+
"step": 195000,
|
| 2123 |
+
"training_step_in_ms": 966.2778741791844
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 80.61,
|
| 2127 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5363166332244873,
|
| 2128 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.2732828183368321,
|
| 2129 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.105,
|
| 2130 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.243,
|
| 2131 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.32,
|
| 2132 |
+
"step": 195000
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"_prepare_inputs_in_ms": 4.114095866680145,
|
| 2136 |
+
"compute_loss_in_ms": 267.9977478161454,
|
| 2137 |
+
"epoch": 81.03,
|
| 2138 |
+
"learning_rate/full": 4.3161571727299287e-07,
|
| 2139 |
+
"loss": 3.5457,
|
| 2140 |
+
"step": 196000,
|
| 2141 |
+
"training_step_in_ms": 963.1876187734306
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"_prepare_inputs_in_ms": 4.132705166935921,
|
| 2145 |
+
"compute_loss_in_ms": 268.0948423668742,
|
| 2146 |
+
"epoch": 81.44,
|
| 2147 |
+
"learning_rate/full": 2.4980400999989885e-07,
|
| 2148 |
+
"loss": 3.5462,
|
| 2149 |
+
"step": 197000,
|
| 2150 |
+
"training_step_in_ms": 967.0936130546033
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"_prepare_inputs_in_ms": 4.152493238449097,
|
| 2154 |
+
"compute_loss_in_ms": 268.8196250721812,
|
| 2155 |
+
"epoch": 81.85,
|
| 2156 |
+
"learning_rate/full": 1.174841397763915e-07,
|
| 2157 |
+
"loss": 3.5462,
|
| 2158 |
+
"step": 198000,
|
| 2159 |
+
"training_step_in_ms": 964.0161675550044
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"_prepare_inputs_in_ms": 4.15958097204566,
|
| 2163 |
+
"compute_loss_in_ms": 268.9542033970356,
|
| 2164 |
+
"epoch": 82.27,
|
| 2165 |
+
"learning_rate/full": 3.4482289435100457e-08,
|
| 2166 |
+
"loss": 3.5489,
|
| 2167 |
+
"step": 199000,
|
| 2168 |
+
"training_step_in_ms": 964.1794747672975
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"_prepare_inputs_in_ms": 4.161274570971727,
|
| 2172 |
+
"compute_loss_in_ms": 268.2289356328547,
|
| 2173 |
+
"epoch": 82.68,
|
| 2174 |
+
"learning_rate/full": 8.099638465708381e-10,
|
| 2175 |
+
"loss": 3.5503,
|
| 2176 |
+
"step": 200000,
|
| 2177 |
+
"training_step_in_ms": 966.4167955368757
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 82.68,
|
| 2181 |
+
"eval_visual_genome-densecap-local-densecap-test_loss": 3.5363588333129883,
|
| 2182 |
+
"eval_visual_genome-densecap-local-densecap-test_meteor": 0.27337310510486335,
|
| 2183 |
+
"eval_visual_genome-densecap-local-densecap-test_runtime": 78.3968,
|
| 2184 |
+
"eval_visual_genome-densecap-local-densecap-test_samples_per_second": 10.204,
|
| 2185 |
+
"eval_visual_genome-densecap-local-densecap-test_steps_per_second": 0.319,
|
| 2186 |
+
"step": 200000
|
| 2187 |
+
}
|
| 2188 |
+
],
|
| 2189 |
+
"max_steps": 200000,
|
| 2190 |
+
"num_train_epochs": 83,
|
| 2191 |
+
"total_flos": 1.6535680980503157e+23,
|
| 2192 |
+
"trial_name": null,
|
| 2193 |
+
"trial_params": null
|
| 2194 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0618515c8c1141dab007214e96ce45520da8c933c71b54ce4d7360861a6948fb
|
| 3 |
+
size 5240
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
zero_to_fp32.py
ADDED
|
@@ -0,0 +1,587 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#!/usr/bin/env python
|
| 2 |
+
|
| 3 |
+
# Copyright (c) Microsoft Corporation.
|
| 4 |
+
# SPDX-License-Identifier: Apache-2.0
|
| 5 |
+
|
| 6 |
+
# DeepSpeed Team
|
| 7 |
+
|
| 8 |
+
# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
|
| 9 |
+
# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
|
| 10 |
+
# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
|
| 11 |
+
# application.
|
| 12 |
+
#
|
| 13 |
+
# example: python zero_to_fp32.py . pytorch_model.bin
|
| 14 |
+
|
| 15 |
+
import argparse
|
| 16 |
+
import torch
|
| 17 |
+
import glob
|
| 18 |
+
import math
|
| 19 |
+
import os
|
| 20 |
+
import re
|
| 21 |
+
from collections import OrderedDict
|
| 22 |
+
from dataclasses import dataclass
|
| 23 |
+
|
| 24 |
+
# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
|
| 25 |
+
# DeepSpeed data structures it has to be available in the current python environment.
|
| 26 |
+
from deepspeed.utils import logger
|
| 27 |
+
from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
|
| 28 |
+
FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
|
| 29 |
+
FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
@dataclass
|
| 33 |
+
class zero_model_state:
|
| 34 |
+
buffers: dict()
|
| 35 |
+
param_shapes: dict()
|
| 36 |
+
shared_params: list
|
| 37 |
+
ds_version: int
|
| 38 |
+
frozen_param_shapes: dict()
|
| 39 |
+
frozen_param_fragments: dict()
|
| 40 |
+
|
| 41 |
+
|
| 42 |
+
debug = 0
|
| 43 |
+
|
| 44 |
+
# load to cpu
|
| 45 |
+
device = torch.device('cpu')
|
| 46 |
+
|
| 47 |
+
|
| 48 |
+
def atoi(text):
|
| 49 |
+
return int(text) if text.isdigit() else text
|
| 50 |
+
|
| 51 |
+
|
| 52 |
+
def natural_keys(text):
|
| 53 |
+
'''
|
| 54 |
+
alist.sort(key=natural_keys) sorts in human order
|
| 55 |
+
http://nedbatchelder.com/blog/200712/human_sorting.html
|
| 56 |
+
(See Toothy's implementation in the comments)
|
| 57 |
+
'''
|
| 58 |
+
return [atoi(c) for c in re.split(r'(\d+)', text)]
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
def get_model_state_file(checkpoint_dir, zero_stage):
|
| 62 |
+
if not os.path.isdir(checkpoint_dir):
|
| 63 |
+
raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
|
| 64 |
+
|
| 65 |
+
# there should be only one file
|
| 66 |
+
if zero_stage <= 2:
|
| 67 |
+
file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
|
| 68 |
+
elif zero_stage == 3:
|
| 69 |
+
file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
|
| 70 |
+
|
| 71 |
+
if not os.path.exists(file):
|
| 72 |
+
raise FileNotFoundError(f"can't find model states file at '{file}'")
|
| 73 |
+
|
| 74 |
+
return file
|
| 75 |
+
|
| 76 |
+
|
| 77 |
+
def get_checkpoint_files(checkpoint_dir, glob_pattern):
|
| 78 |
+
# XXX: need to test that this simple glob rule works for multi-node setup too
|
| 79 |
+
ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
|
| 80 |
+
|
| 81 |
+
if len(ckpt_files) == 0:
|
| 82 |
+
raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
|
| 83 |
+
|
| 84 |
+
return ckpt_files
|
| 85 |
+
|
| 86 |
+
|
| 87 |
+
def get_optim_files(checkpoint_dir):
|
| 88 |
+
return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
|
| 89 |
+
|
| 90 |
+
|
| 91 |
+
def get_model_state_files(checkpoint_dir):
|
| 92 |
+
return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
|
| 93 |
+
|
| 94 |
+
|
| 95 |
+
def parse_model_states(files):
|
| 96 |
+
zero_model_states = []
|
| 97 |
+
for file in files:
|
| 98 |
+
state_dict = torch.load(file, map_location=device)
|
| 99 |
+
|
| 100 |
+
if BUFFER_NAMES not in state_dict:
|
| 101 |
+
raise ValueError(f"{file} is not a model state checkpoint")
|
| 102 |
+
buffer_names = state_dict[BUFFER_NAMES]
|
| 103 |
+
if debug:
|
| 104 |
+
print("Found buffers:", buffer_names)
|
| 105 |
+
|
| 106 |
+
# recover just the buffers while restoring them to fp32 if they were saved in fp16
|
| 107 |
+
buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
|
| 108 |
+
param_shapes = state_dict[PARAM_SHAPES]
|
| 109 |
+
|
| 110 |
+
# collect parameters that are included in param_shapes
|
| 111 |
+
param_names = []
|
| 112 |
+
for s in param_shapes:
|
| 113 |
+
for name in s.keys():
|
| 114 |
+
param_names.append(name)
|
| 115 |
+
|
| 116 |
+
# update with frozen parameters
|
| 117 |
+
frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
|
| 118 |
+
if frozen_param_shapes is not None:
|
| 119 |
+
if debug:
|
| 120 |
+
print(f"Found frozen_param_shapes: {frozen_param_shapes}")
|
| 121 |
+
param_names += list(frozen_param_shapes.keys())
|
| 122 |
+
|
| 123 |
+
# handle shared params
|
| 124 |
+
shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
|
| 125 |
+
|
| 126 |
+
ds_version = state_dict.get(DS_VERSION, None)
|
| 127 |
+
|
| 128 |
+
frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
|
| 129 |
+
|
| 130 |
+
z_model_state = zero_model_state(buffers=buffers,
|
| 131 |
+
param_shapes=param_shapes,
|
| 132 |
+
shared_params=shared_params,
|
| 133 |
+
ds_version=ds_version,
|
| 134 |
+
frozen_param_shapes=frozen_param_shapes,
|
| 135 |
+
frozen_param_fragments=frozen_param_fragments)
|
| 136 |
+
zero_model_states.append(z_model_state)
|
| 137 |
+
|
| 138 |
+
return zero_model_states
|
| 139 |
+
|
| 140 |
+
|
| 141 |
+
def parse_optim_states(files, ds_checkpoint_dir):
|
| 142 |
+
|
| 143 |
+
total_files = len(files)
|
| 144 |
+
state_dicts = []
|
| 145 |
+
for f in files:
|
| 146 |
+
state_dict = torch.load(f, map_location=device)
|
| 147 |
+
# immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
|
| 148 |
+
# and also handle the case where it was already removed by another helper script
|
| 149 |
+
state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
|
| 150 |
+
state_dicts.append(state_dict)
|
| 151 |
+
|
| 152 |
+
if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
|
| 153 |
+
raise ValueError(f"{files[0]} is not a zero checkpoint")
|
| 154 |
+
zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
|
| 155 |
+
world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
|
| 156 |
+
|
| 157 |
+
# For ZeRO-2 each param group can have different partition_count as data parallelism for expert
|
| 158 |
+
# parameters can be different from data parallelism for non-expert parameters. So we can just
|
| 159 |
+
# use the max of the partition_count to get the dp world_size.
|
| 160 |
+
|
| 161 |
+
if type(world_size) is list:
|
| 162 |
+
world_size = max(world_size)
|
| 163 |
+
|
| 164 |
+
if world_size != total_files:
|
| 165 |
+
raise ValueError(
|
| 166 |
+
f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
|
| 167 |
+
"Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
|
| 168 |
+
)
|
| 169 |
+
|
| 170 |
+
# the groups are named differently in each stage
|
| 171 |
+
if zero_stage <= 2:
|
| 172 |
+
fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
|
| 173 |
+
elif zero_stage == 3:
|
| 174 |
+
fp32_groups_key = FP32_FLAT_GROUPS
|
| 175 |
+
else:
|
| 176 |
+
raise ValueError(f"unknown zero stage {zero_stage}")
|
| 177 |
+
|
| 178 |
+
if zero_stage <= 2:
|
| 179 |
+
fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
|
| 180 |
+
elif zero_stage == 3:
|
| 181 |
+
# if there is more than one param group, there will be multiple flattened tensors - one
|
| 182 |
+
# flattened tensor per group - for simplicity merge them into a single tensor
|
| 183 |
+
#
|
| 184 |
+
# XXX: could make the script more memory efficient for when there are multiple groups - it
|
| 185 |
+
# will require matching the sub-lists of param_shapes for each param group flattened tensor
|
| 186 |
+
|
| 187 |
+
fp32_flat_groups = [
|
| 188 |
+
torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
|
| 189 |
+
]
|
| 190 |
+
|
| 191 |
+
return zero_stage, world_size, fp32_flat_groups
|
| 192 |
+
|
| 193 |
+
|
| 194 |
+
def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir):
|
| 195 |
+
"""
|
| 196 |
+
Returns fp32 state_dict reconstructed from ds checkpoint
|
| 197 |
+
|
| 198 |
+
Args:
|
| 199 |
+
- ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
|
| 200 |
+
|
| 201 |
+
"""
|
| 202 |
+
print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
|
| 203 |
+
|
| 204 |
+
optim_files = get_optim_files(ds_checkpoint_dir)
|
| 205 |
+
zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
|
| 206 |
+
print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
|
| 207 |
+
|
| 208 |
+
model_files = get_model_state_files(ds_checkpoint_dir)
|
| 209 |
+
|
| 210 |
+
zero_model_states = parse_model_states(model_files)
|
| 211 |
+
print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
|
| 212 |
+
|
| 213 |
+
if zero_stage <= 2:
|
| 214 |
+
return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states)
|
| 215 |
+
elif zero_stage == 3:
|
| 216 |
+
return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states)
|
| 217 |
+
|
| 218 |
+
|
| 219 |
+
def _zero2_merge_frozen_params(state_dict, zero_model_states):
|
| 220 |
+
if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
|
| 221 |
+
return
|
| 222 |
+
|
| 223 |
+
frozen_param_shapes = zero_model_states[0].frozen_param_shapes
|
| 224 |
+
frozen_param_fragments = zero_model_states[0].frozen_param_fragments
|
| 225 |
+
|
| 226 |
+
if debug:
|
| 227 |
+
num_elem = sum(s.numel() for s in frozen_param_shapes.values())
|
| 228 |
+
print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
|
| 229 |
+
|
| 230 |
+
wanted_params = len(frozen_param_shapes)
|
| 231 |
+
wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
|
| 232 |
+
avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
|
| 233 |
+
print(f'Frozen params: Have {avail_numel} numels to process.')
|
| 234 |
+
print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
|
| 235 |
+
|
| 236 |
+
total_params = 0
|
| 237 |
+
total_numel = 0
|
| 238 |
+
for name, shape in frozen_param_shapes.items():
|
| 239 |
+
total_params += 1
|
| 240 |
+
unpartitioned_numel = shape.numel()
|
| 241 |
+
total_numel += unpartitioned_numel
|
| 242 |
+
|
| 243 |
+
state_dict[name] = frozen_param_fragments[name]
|
| 244 |
+
|
| 245 |
+
if debug:
|
| 246 |
+
print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
|
| 247 |
+
|
| 248 |
+
print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
|
| 249 |
+
|
| 250 |
+
|
| 251 |
+
def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
|
| 252 |
+
param_shapes = zero_model_states[0].param_shapes
|
| 253 |
+
|
| 254 |
+
# Reconstruction protocol:
|
| 255 |
+
#
|
| 256 |
+
# XXX: document this
|
| 257 |
+
|
| 258 |
+
if debug:
|
| 259 |
+
for i in range(world_size):
|
| 260 |
+
for j in range(len(fp32_flat_groups[0])):
|
| 261 |
+
print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
|
| 262 |
+
|
| 263 |
+
# XXX: memory usage doubles here (zero2)
|
| 264 |
+
num_param_groups = len(fp32_flat_groups[0])
|
| 265 |
+
merged_single_partition_of_fp32_groups = []
|
| 266 |
+
for i in range(num_param_groups):
|
| 267 |
+
merged_partitions = [sd[i] for sd in fp32_flat_groups]
|
| 268 |
+
full_single_fp32_vector = torch.cat(merged_partitions, 0)
|
| 269 |
+
merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
|
| 270 |
+
avail_numel = sum(
|
| 271 |
+
[full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
|
| 272 |
+
|
| 273 |
+
if debug:
|
| 274 |
+
wanted_params = sum([len(shapes) for shapes in param_shapes])
|
| 275 |
+
wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
|
| 276 |
+
# not asserting if there is a mismatch due to possible padding
|
| 277 |
+
print(f"Have {avail_numel} numels to process.")
|
| 278 |
+
print(f"Need {wanted_numel} numels in {wanted_params} params.")
|
| 279 |
+
|
| 280 |
+
# params
|
| 281 |
+
# XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
|
| 282 |
+
# out-of-core computing solution
|
| 283 |
+
total_numel = 0
|
| 284 |
+
total_params = 0
|
| 285 |
+
for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
|
| 286 |
+
offset = 0
|
| 287 |
+
avail_numel = full_single_fp32_vector.numel()
|
| 288 |
+
for name, shape in shapes.items():
|
| 289 |
+
|
| 290 |
+
unpartitioned_numel = shape.numel()
|
| 291 |
+
total_numel += unpartitioned_numel
|
| 292 |
+
total_params += 1
|
| 293 |
+
|
| 294 |
+
if debug:
|
| 295 |
+
print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
|
| 296 |
+
state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
|
| 297 |
+
offset += unpartitioned_numel
|
| 298 |
+
|
| 299 |
+
# Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
|
| 300 |
+
# avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
|
| 301 |
+
# paddings performed in the code it's almost impossible to predict the exact numbers w/o the
|
| 302 |
+
# live optimizer object, so we are checking that the numbers are within the right range
|
| 303 |
+
align_to = 2 * world_size
|
| 304 |
+
|
| 305 |
+
def zero2_align(x):
|
| 306 |
+
return align_to * math.ceil(x / align_to)
|
| 307 |
+
|
| 308 |
+
if debug:
|
| 309 |
+
print(f"original offset={offset}, avail_numel={avail_numel}")
|
| 310 |
+
|
| 311 |
+
offset = zero2_align(offset)
|
| 312 |
+
avail_numel = zero2_align(avail_numel)
|
| 313 |
+
|
| 314 |
+
if debug:
|
| 315 |
+
print(f"aligned offset={offset}, avail_numel={avail_numel}")
|
| 316 |
+
|
| 317 |
+
# Sanity check
|
| 318 |
+
if offset != avail_numel:
|
| 319 |
+
raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
|
| 320 |
+
|
| 321 |
+
print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
|
| 322 |
+
|
| 323 |
+
|
| 324 |
+
def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states):
|
| 325 |
+
state_dict = OrderedDict()
|
| 326 |
+
|
| 327 |
+
# buffers
|
| 328 |
+
buffers = zero_model_states[0].buffers
|
| 329 |
+
state_dict.update(buffers)
|
| 330 |
+
if debug:
|
| 331 |
+
print(f"added {len(buffers)} buffers")
|
| 332 |
+
|
| 333 |
+
_zero2_merge_frozen_params(state_dict, zero_model_states)
|
| 334 |
+
|
| 335 |
+
_zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
|
| 336 |
+
|
| 337 |
+
# recover shared parameters
|
| 338 |
+
for pair in zero_model_states[0].shared_params:
|
| 339 |
+
if pair[1] in state_dict:
|
| 340 |
+
state_dict[pair[0]] = state_dict[pair[1]]
|
| 341 |
+
|
| 342 |
+
return state_dict
|
| 343 |
+
|
| 344 |
+
|
| 345 |
+
def zero3_partitioned_param_info(unpartitioned_numel, world_size):
|
| 346 |
+
remainder = unpartitioned_numel % world_size
|
| 347 |
+
padding_numel = (world_size - remainder) if remainder else 0
|
| 348 |
+
partitioned_numel = math.ceil(unpartitioned_numel / world_size)
|
| 349 |
+
return partitioned_numel, padding_numel
|
| 350 |
+
|
| 351 |
+
|
| 352 |
+
def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
|
| 353 |
+
if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
|
| 354 |
+
return
|
| 355 |
+
|
| 356 |
+
if debug:
|
| 357 |
+
for i in range(world_size):
|
| 358 |
+
num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
|
| 359 |
+
print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
|
| 360 |
+
|
| 361 |
+
frozen_param_shapes = zero_model_states[0].frozen_param_shapes
|
| 362 |
+
wanted_params = len(frozen_param_shapes)
|
| 363 |
+
wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
|
| 364 |
+
avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
|
| 365 |
+
print(f'Frozen params: Have {avail_numel} numels to process.')
|
| 366 |
+
print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
|
| 367 |
+
|
| 368 |
+
total_params = 0
|
| 369 |
+
total_numel = 0
|
| 370 |
+
for name, shape in zero_model_states[0].frozen_param_shapes.items():
|
| 371 |
+
total_params += 1
|
| 372 |
+
unpartitioned_numel = shape.numel()
|
| 373 |
+
total_numel += unpartitioned_numel
|
| 374 |
+
|
| 375 |
+
param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
|
| 376 |
+
state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
|
| 377 |
+
|
| 378 |
+
partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
|
| 379 |
+
|
| 380 |
+
if debug:
|
| 381 |
+
print(
|
| 382 |
+
f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
|
| 383 |
+
)
|
| 384 |
+
|
| 385 |
+
print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
|
| 386 |
+
|
| 387 |
+
|
| 388 |
+
def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
|
| 389 |
+
param_shapes = zero_model_states[0].param_shapes
|
| 390 |
+
avail_numel = fp32_flat_groups[0].numel() * world_size
|
| 391 |
+
# Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
|
| 392 |
+
# param, re-consolidating each param, while dealing with padding if any
|
| 393 |
+
|
| 394 |
+
# merge list of dicts, preserving order
|
| 395 |
+
param_shapes = {k: v for d in param_shapes for k, v in d.items()}
|
| 396 |
+
|
| 397 |
+
if debug:
|
| 398 |
+
for i in range(world_size):
|
| 399 |
+
print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
|
| 400 |
+
|
| 401 |
+
wanted_params = len(param_shapes)
|
| 402 |
+
wanted_numel = sum(shape.numel() for shape in param_shapes.values())
|
| 403 |
+
# not asserting if there is a mismatch due to possible padding
|
| 404 |
+
avail_numel = fp32_flat_groups[0].numel() * world_size
|
| 405 |
+
print(f"Trainable params: Have {avail_numel} numels to process.")
|
| 406 |
+
print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
|
| 407 |
+
|
| 408 |
+
# params
|
| 409 |
+
# XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
|
| 410 |
+
# out-of-core computing solution
|
| 411 |
+
offset = 0
|
| 412 |
+
total_numel = 0
|
| 413 |
+
total_params = 0
|
| 414 |
+
for name, shape in param_shapes.items():
|
| 415 |
+
|
| 416 |
+
unpartitioned_numel = shape.numel()
|
| 417 |
+
total_numel += unpartitioned_numel
|
| 418 |
+
total_params += 1
|
| 419 |
+
|
| 420 |
+
partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
|
| 421 |
+
|
| 422 |
+
if debug:
|
| 423 |
+
print(
|
| 424 |
+
f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
|
| 425 |
+
)
|
| 426 |
+
|
| 427 |
+
# XXX: memory usage doubles here
|
| 428 |
+
state_dict[name] = torch.cat(
|
| 429 |
+
tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
|
| 430 |
+
0).narrow(0, 0, unpartitioned_numel).view(shape)
|
| 431 |
+
offset += partitioned_numel
|
| 432 |
+
|
| 433 |
+
offset *= world_size
|
| 434 |
+
|
| 435 |
+
# Sanity check
|
| 436 |
+
if offset != avail_numel:
|
| 437 |
+
raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
|
| 438 |
+
|
| 439 |
+
print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
|
| 440 |
+
|
| 441 |
+
|
| 442 |
+
def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states):
|
| 443 |
+
state_dict = OrderedDict()
|
| 444 |
+
|
| 445 |
+
# buffers
|
| 446 |
+
buffers = zero_model_states[0].buffers
|
| 447 |
+
state_dict.update(buffers)
|
| 448 |
+
if debug:
|
| 449 |
+
print(f"added {len(buffers)} buffers")
|
| 450 |
+
|
| 451 |
+
_zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
|
| 452 |
+
|
| 453 |
+
_zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
|
| 454 |
+
|
| 455 |
+
# recover shared parameters
|
| 456 |
+
for pair in zero_model_states[0].shared_params:
|
| 457 |
+
if pair[1] in state_dict:
|
| 458 |
+
state_dict[pair[0]] = state_dict[pair[1]]
|
| 459 |
+
|
| 460 |
+
return state_dict
|
| 461 |
+
|
| 462 |
+
|
| 463 |
+
def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None):
|
| 464 |
+
"""
|
| 465 |
+
Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
|
| 466 |
+
``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
|
| 467 |
+
via a model hub.
|
| 468 |
+
|
| 469 |
+
Args:
|
| 470 |
+
- ``checkpoint_dir``: path to the desired checkpoint folder
|
| 471 |
+
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
|
| 472 |
+
|
| 473 |
+
Returns:
|
| 474 |
+
- pytorch ``state_dict``
|
| 475 |
+
|
| 476 |
+
Note: this approach may not work if your application doesn't have sufficient free CPU memory and
|
| 477 |
+
you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
|
| 478 |
+
the checkpoint.
|
| 479 |
+
|
| 480 |
+
A typical usage might be ::
|
| 481 |
+
|
| 482 |
+
from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
|
| 483 |
+
# do the training and checkpoint saving
|
| 484 |
+
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
|
| 485 |
+
model = model.cpu() # move to cpu
|
| 486 |
+
model.load_state_dict(state_dict)
|
| 487 |
+
# submit to model hub or save the model to share with others
|
| 488 |
+
|
| 489 |
+
In this example the ``model`` will no longer be usable in the deepspeed context of the same
|
| 490 |
+
application. i.e. you will need to re-initialize the deepspeed engine, since
|
| 491 |
+
``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
|
| 492 |
+
|
| 493 |
+
If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
|
| 494 |
+
|
| 495 |
+
"""
|
| 496 |
+
if tag is None:
|
| 497 |
+
latest_path = os.path.join(checkpoint_dir, 'latest')
|
| 498 |
+
if os.path.isfile(latest_path):
|
| 499 |
+
with open(latest_path, 'r') as fd:
|
| 500 |
+
tag = fd.read().strip()
|
| 501 |
+
else:
|
| 502 |
+
raise ValueError(f"Unable to find 'latest' file at {latest_path}")
|
| 503 |
+
|
| 504 |
+
ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
|
| 505 |
+
|
| 506 |
+
if not os.path.isdir(ds_checkpoint_dir):
|
| 507 |
+
raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
|
| 508 |
+
|
| 509 |
+
return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir)
|
| 510 |
+
|
| 511 |
+
|
| 512 |
+
def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None):
|
| 513 |
+
"""
|
| 514 |
+
Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
|
| 515 |
+
loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
|
| 516 |
+
|
| 517 |
+
Args:
|
| 518 |
+
- ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
|
| 519 |
+
- ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
|
| 520 |
+
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
|
| 521 |
+
"""
|
| 522 |
+
|
| 523 |
+
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
|
| 524 |
+
print(f"Saving fp32 state dict to {output_file}")
|
| 525 |
+
torch.save(state_dict, output_file)
|
| 526 |
+
|
| 527 |
+
|
| 528 |
+
def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
|
| 529 |
+
"""
|
| 530 |
+
1. Put the provided model to cpu
|
| 531 |
+
2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
|
| 532 |
+
3. Load it into the provided model
|
| 533 |
+
|
| 534 |
+
Args:
|
| 535 |
+
- ``model``: the model object to update
|
| 536 |
+
- ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
|
| 537 |
+
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
|
| 538 |
+
|
| 539 |
+
Returns:
|
| 540 |
+
- ``model`: modified model
|
| 541 |
+
|
| 542 |
+
Make sure you have plenty of CPU memory available before you call this function. If you don't
|
| 543 |
+
have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
|
| 544 |
+
conveniently placed for you in the checkpoint folder.
|
| 545 |
+
|
| 546 |
+
A typical usage might be ::
|
| 547 |
+
|
| 548 |
+
from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
|
| 549 |
+
model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
|
| 550 |
+
# submit to model hub or save the model to share with others
|
| 551 |
+
|
| 552 |
+
Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
|
| 553 |
+
of the same application. i.e. you will need to re-initialize the deepspeed engine, since
|
| 554 |
+
``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
|
| 555 |
+
|
| 556 |
+
"""
|
| 557 |
+
logger.info(f"Extracting fp32 weights")
|
| 558 |
+
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
|
| 559 |
+
|
| 560 |
+
logger.info(f"Overwriting model with fp32 weights")
|
| 561 |
+
model = model.cpu()
|
| 562 |
+
model.load_state_dict(state_dict, strict=False)
|
| 563 |
+
|
| 564 |
+
return model
|
| 565 |
+
|
| 566 |
+
|
| 567 |
+
if __name__ == "__main__":
|
| 568 |
+
|
| 569 |
+
parser = argparse.ArgumentParser()
|
| 570 |
+
parser.add_argument("checkpoint_dir",
|
| 571 |
+
type=str,
|
| 572 |
+
help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
|
| 573 |
+
parser.add_argument(
|
| 574 |
+
"output_file",
|
| 575 |
+
type=str,
|
| 576 |
+
help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
|
| 577 |
+
parser.add_argument("-t",
|
| 578 |
+
"--tag",
|
| 579 |
+
type=str,
|
| 580 |
+
default=None,
|
| 581 |
+
help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
|
| 582 |
+
parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
|
| 583 |
+
args = parser.parse_args()
|
| 584 |
+
|
| 585 |
+
debug = args.debug
|
| 586 |
+
|
| 587 |
+
convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, args.output_file, tag=args.tag)
|