End of training

README.md

@@ -2,10 +2,13 @@
 license: apache-2.0
 library_name: peft
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
 base_model: alignment-handbook/zephyr-7b-sft-full
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: zephyr-7b-dpo-lora
   results: []
@@ -16,17 +19,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-dpo-lora
 
-This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the None dataset.
+This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.5893
-- Rewards/chosen: -0.2740
-- Rewards/rejected: -0.6023
-- Rewards/accuracies: 0.7025
-- Rewards/margins: 0.3283
-- Logps/rejected: -321.6666
-- Logps/chosen: -310.1333
-- Logits/rejected: -2.7525
-- Logits/chosen: -2.7742
+- Loss: 0.5894
+- Rewards/chosen: -0.2738
+- Rewards/rejected: -0.6020
+- Rewards/accuracies: 0.7035
+- Rewards/margins: 0.3282
+- Logps/rejected: -321.6407
+- Logps/chosen: -310.1199
+- Logits/rejected: -2.7529
+- Logits/chosen: -2.7746
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -2.7746472358703613,
+    "eval_logits/rejected": -2.752934455871582,
+    "eval_logps/chosen": -310.119873046875,
+    "eval_logps/rejected": -321.6407165527344,
+    "eval_loss": 0.5893968343734741,
+    "eval_rewards/accuracies": 0.703499972820282,
+    "eval_rewards/chosen": -0.27382245659828186,
+    "eval_rewards/margins": 0.32820531725883484,
+    "eval_rewards/rejected": -0.6020277142524719,
+    "eval_runtime": 692.2285,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 2.889,
+    "eval_steps_per_second": 0.361,
     "total_flos": 0.0,
     "train_loss": 0.6164219083351729,
     "train_runtime": 73481.1174,

config.json

@@ -0,0 +1,25 @@
+{
+  "_name_or_path": "alignment-handbook/zephyr-7b-sft-full",
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 10000.0,
+  "sliding_window": 4096,
+  "tie_word_embeddings": false,
+  "transformers_version": "4.40.0",
+  "use_cache": true,
+  "vocab_size": 32000
+}

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 1.0,
+    "eval_logits/chosen": -2.7746472358703613,
+    "eval_logits/rejected": -2.752934455871582,
+    "eval_logps/chosen": -310.119873046875,
+    "eval_logps/rejected": -321.6407165527344,
+    "eval_loss": 0.5893968343734741,
+    "eval_rewards/accuracies": 0.703499972820282,
+    "eval_rewards/chosen": -0.27382245659828186,
+    "eval_rewards/margins": 0.32820531725883484,
+    "eval_rewards/rejected": -0.6020277142524719,
+    "eval_runtime": 692.2285,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 2.889,
+    "eval_steps_per_second": 0.361
+}