uploading better model with 760 mln params

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "sberbank-ai/rugpt3small_based_on_gpt2",
+  "_name_or_path": "sberbank-ai/rugpt3large_based_on_gpt2",
   "activation_function": "gelu_new",
   "architectures": [
     "GPT2LMHeadModel"
@@ -13,10 +13,10 @@
   "layer_norm_epsilon": 1e-05,
   "model_type": "gpt2",
   "n_ctx": 2048,
-  "n_embd": 768,
-  "n_head": 12,
+  "n_embd": 1536,
+  "n_head": 16,
   "n_inner": null,
-  "n_layer": 12,
+  "n_layer": 24,
   "n_positions": 2048,
   "reorder_and_upcast_attn": false,
   "resid_pdrop": 0.1,
@@ -28,7 +28,7 @@
   "summary_type": "cls_index",
   "summary_use_proj": true,
   "torch_dtype": "float32",
-  "transformers_version": "4.21.3",
+  "transformers_version": "4.21.1",
   "use_cache": true,
   "vocab_size": 50258
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:98ae88bfbed22ba3fb913bec46ae4a206825bf67b240af7057f88315a3090b90
-size 551296803
+oid sha256:5207dc71dfa8d825fbab940864930b2c9b7e0e43b28a0cd124522a9185413303
+size 3141986147

tokenizer_config.json

@@ -18,7 +18,7 @@
     "single_word": false
   },
   "errors": "replace",
-  "name_or_path": "sberbank-ai/rugpt3small_based_on_gpt2",
+  "name_or_path": "sberbank-ai/rugpt3large_based_on_gpt2",
   "pad_token": null,
   "special_tokens_map_file": null,
   "tokenizer_class": "GPT2Tokenizer",