End of training

README.md

@@ -1,10 +1,13 @@
 ---
 library_name: peft
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
 base_model: data/zephyr-7b-sft-qlora-merged
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: zephyr-7b-dpo-qlora
   results: []
@@ -16,17 +19,17 @@ should probably proofread and complete it, then remove this comment. -->
 [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/statking/huggingface/runs/qxp2vmm7)
 # zephyr-7b-dpo-qlora
 
-This model was trained from scratch on an unknown dataset.
+This model is a fine-tuned version of [data/zephyr-7b-sft-qlora-merged](https://huggingface.co/data/zephyr-7b-sft-qlora-merged) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4933
-- Rewards/chosen: -2.1713
-- Rewards/rejected: -3.1801
-- Rewards/accuracies: 0.7738
-- Rewards/margins: 1.0088
-- Logps/rejected: -564.8470
-- Logps/chosen: -483.4024
-- Logits/rejected: -1.4105
-- Logits/chosen: -1.4778
+- Loss: 0.4935
+- Rewards/chosen: -2.1676
+- Rewards/rejected: -3.1735
+- Rewards/accuracies: 0.7698
+- Rewards/margins: 1.0059
+- Logps/rejected: -564.1859
+- Logps/chosen: -483.0326
+- Logits/rejected: -1.4139
+- Logits/chosen: -1.4811
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 0.9997382884061764,
+    "eval_logits/chosen": -1.4810994863510132,
+    "eval_logits/rejected": -1.413936734199524,
+    "eval_logps/chosen": -483.0326232910156,
+    "eval_logps/rejected": -564.1859130859375,
+    "eval_loss": 0.4935336709022522,
+    "eval_rewards/accuracies": 0.7698412537574768,
+    "eval_rewards/chosen": -2.1675848960876465,
+    "eval_rewards/margins": 1.0059263706207275,
+    "eval_rewards/rejected": -3.173511266708374,
+    "eval_runtime": 276.8986,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 7.223,
+    "eval_steps_per_second": 0.228,
     "total_flos": 0.0,
     "train_loss": 0.5333428270529702,
     "train_runtime": 30179.1373,

config.json

@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "data/zephyr-7b-sft-qlora-merged",
+  "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,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.41.0.dev0",
+  "use_cache": true,
+  "vocab_size": 32000
+}

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 0.9997382884061764,
+    "eval_logits/chosen": -1.4810994863510132,
+    "eval_logits/rejected": -1.413936734199524,
+    "eval_logps/chosen": -483.0326232910156,
+    "eval_logps/rejected": -564.1859130859375,
+    "eval_loss": 0.4935336709022522,
+    "eval_rewards/accuracies": 0.7698412537574768,
+    "eval_rewards/chosen": -2.1675848960876465,
+    "eval_rewards/margins": 1.0059263706207275,
+    "eval_rewards/rejected": -3.173511266708374,
+    "eval_runtime": 276.8986,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 7.223,
+    "eval_steps_per_second": 0.228
+}