End of training

README.md

@@ -2,10 +2,13 @@
 license: llama3
 library_name: peft
 tags:
+- alignment-handbook
 - trl
 - orpo
 - generated_from_trainer
 base_model: meta-llama/Meta-Llama-3-8B
+datasets:
+- mlabonne/orpo-dpo-mix-40k
 model-index:
 - name: llama-3-orpo-qlora
   results: []
@@ -16,7 +19,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # llama-3-orpo-qlora
 
-This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) on an unknown dataset.
+This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) on the mlabonne/orpo-dpo-mix-40k dataset.
 It achieves the following results on the evaluation set:
 - Loss: 1.0581
 - Rewards/chosen: -0.0823

all_results.json

@@ -1,5 +1,21 @@
 {
     "epoch": 4.999237456153729,
+    "eval_log_odds_chosen": 2.101734161376953,
+    "eval_log_odds_ratio": -0.39467301964759827,
+    "eval_logits/chosen": -0.9354549646377563,
+    "eval_logits/rejected": -1.034712791442871,
+    "eval_logps/chosen": -0.8229629993438721,
+    "eval_logps/rejected": -2.495842695236206,
+    "eval_loss": 1.0581330060958862,
+    "eval_nll_loss": 1.0625019073486328,
+    "eval_rewards/accuracies": 0.7878788113594055,
+    "eval_rewards/chosen": -0.08229630440473557,
+    "eval_rewards/margins": 0.16728799045085907,
+    "eval_rewards/rejected": -0.24958431720733643,
+    "eval_runtime": 394.4103,
+    "eval_samples": 393,
+    "eval_samples_per_second": 0.996,
+    "eval_steps_per_second": 0.167,
     "total_flos": 0.0,
     "train_loss": 1.1049469483633314,
     "train_runtime": 586416.3764,

config.json

@@ -0,0 +1,30 @@
+{
+  "_name_or_path": "meta-llama/Meta-Llama-3-8B",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128256,
+  "eos_token_id": 128257,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 8192,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pad_token_id": 128257,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.41.2",
+  "use_cache": true,
+  "vocab_size": 128258
+}

eval_results.json

@@ -0,0 +1,19 @@
+{
+    "epoch": 4.999237456153729,
+    "eval_log_odds_chosen": 2.101734161376953,
+    "eval_log_odds_ratio": -0.39467301964759827,
+    "eval_logits/chosen": -0.9354549646377563,
+    "eval_logits/rejected": -1.034712791442871,
+    "eval_logps/chosen": -0.8229629993438721,
+    "eval_logps/rejected": -2.495842695236206,
+    "eval_loss": 1.0581330060958862,
+    "eval_nll_loss": 1.0625019073486328,
+    "eval_rewards/accuracies": 0.7878788113594055,
+    "eval_rewards/chosen": -0.08229630440473557,
+    "eval_rewards/margins": 0.16728799045085907,
+    "eval_rewards/rejected": -0.24958431720733643,
+    "eval_runtime": 394.4103,
+    "eval_samples": 393,
+    "eval_samples_per_second": 0.996,
+    "eval_steps_per_second": 0.167
+}