Model save

README.md

@@ -0,0 +1,58 @@
+---
+license: llama2
+base_model: meta-llama/Llama-2-7b-hf
+tags:
+- generated_from_trainer
+model-index:
+- name: Llama-2-7b-dpo-2k
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# Llama-2-7b-dpo-2k
+
+This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on an unknown dataset.
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 5e-07
+- train_batch_size: 4
+- eval_batch_size: 4
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 4
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 32
+- total_eval_batch_size: 16
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: linear
+- lr_scheduler_warmup_ratio: 0.1
+- num_epochs: 1
+
+### Training results
+
+
+
+### Framework versions
+
+- Transformers 4.43.3
+- Pytorch 2.2.2+cu121
+- Datasets 2.20.0
+- Tokenizers 0.19.1

all_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 0.992,
+    "total_flos": 0.0,
+    "train_loss": 0.8413216990809287,
+    "train_runtime": 704.1249,
+    "train_samples": 2000,
+    "train_samples_per_second": 2.84,
+    "train_steps_per_second": 0.088
+}

generation_config.json

@@ -0,0 +1,10 @@
+{
+  "bos_token_id": 1,
+  "do_sample": true,
+  "eos_token_id": 2,
+  "max_length": 4096,
+  "pad_token_id": 0,
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.43.3"
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 0.992,
+    "total_flos": 0.0,
+    "train_loss": 0.8413216990809287,
+    "train_runtime": 704.1249,
+    "train_samples": 2000,
+    "train_samples_per_second": 2.84,
+    "train_steps_per_second": 0.088
+}

trainer_state.json

@@ -0,0 +1,147 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.992,
+  "eval_steps": 200,
+  "global_step": 62,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.016,
+      "grad_norm": 22.11647064104875,
+      "learning_rate": 7.142857142857142e-08,
+      "logits/generated": -0.9303410053253174,
+      "logits/real": -0.6472625732421875,
+      "logps/generated": -159.71011352539062,
+      "logps/real": -190.3052978515625,
+      "loss": 0.8876,
+      "rewards/accuracies": 0.0,
+      "rewards/generated": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/real": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 19.1368380186354,
+      "learning_rate": 4.727272727272727e-07,
+      "logits/generated": -0.7768990993499756,
+      "logits/real": -0.7621282339096069,
+      "logps/generated": -174.13926696777344,
+      "logps/real": -182.75685119628906,
+      "loss": 0.8561,
+      "rewards/accuracies": 0.5416666865348816,
+      "rewards/generated": 0.08834926038980484,
+      "rewards/margins": 0.026768552139401436,
+      "rewards/real": 0.11511781811714172,
+      "step": 10
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 20.383391385284007,
+      "learning_rate": 3.818181818181818e-07,
+      "logits/generated": -0.7519482970237732,
+      "logits/real": -0.7631895542144775,
+      "logps/generated": -163.14199829101562,
+      "logps/real": -164.8223876953125,
+      "loss": 0.8531,
+      "rewards/accuracies": 0.512499988079071,
+      "rewards/generated": 0.46173515915870667,
+      "rewards/margins": -0.009647580794990063,
+      "rewards/real": 0.45208755135536194,
+      "step": 20
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 19.97093382156585,
+      "learning_rate": 2.909090909090909e-07,
+      "logits/generated": -0.8313862085342407,
+      "logits/real": -0.7039824724197388,
+      "logps/generated": -145.91732788085938,
+      "logps/real": -154.5005645751953,
+      "loss": 0.8307,
+      "rewards/accuracies": 0.637499988079071,
+      "rewards/generated": 0.6693741679191589,
+      "rewards/margins": 0.10465053468942642,
+      "rewards/real": 0.7740246653556824,
+      "step": 30
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 20.49095190273365,
+      "learning_rate": 2e-07,
+      "logits/generated": -0.7498478293418884,
+      "logits/real": -0.6850128769874573,
+      "logps/generated": -165.49362182617188,
+      "logps/real": -177.91653442382812,
+      "loss": 0.838,
+      "rewards/accuracies": 0.550000011920929,
+      "rewards/generated": 0.9019988775253296,
+      "rewards/margins": 0.0408473014831543,
+      "rewards/real": 0.9428462982177734,
+      "step": 40
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 20.66774528347473,
+      "learning_rate": 1.0909090909090908e-07,
+      "logits/generated": -0.7961899042129517,
+      "logits/real": -0.621185302734375,
+      "logps/generated": -163.02835083007812,
+      "logps/real": -180.72848510742188,
+      "loss": 0.8404,
+      "rewards/accuracies": 0.6499999761581421,
+      "rewards/generated": 1.0224363803863525,
+      "rewards/margins": 0.09271160513162613,
+      "rewards/real": 1.1151480674743652,
+      "step": 50
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 20.43138637589313,
+      "learning_rate": 1.818181818181818e-08,
+      "logits/generated": -0.6595975160598755,
+      "logits/real": -0.5414831638336182,
+      "logps/generated": -169.04844665527344,
+      "logps/real": -175.93373107910156,
+      "loss": 0.8306,
+      "rewards/accuracies": 0.5625,
+      "rewards/generated": 1.1264116764068604,
+      "rewards/margins": 0.07433578372001648,
+      "rewards/real": 1.2007472515106201,
+      "step": 60
+    },
+    {
+      "epoch": 0.992,
+      "step": 62,
+      "total_flos": 0.0,
+      "train_loss": 0.8413216990809287,
+      "train_runtime": 704.1249,
+      "train_samples_per_second": 2.84,
+      "train_steps_per_second": 0.088
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 62,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 200,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}