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:
+- generation/GPT4
 model-index:
 - name: zephyr-dpop-qlora-gpt4-5e-6-epoch3
   results: []
@@ -16,22 +19,22 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-dpop-qlora-gpt4-5e-6-epoch3
 
-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 generation/GPT4 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 14.5955
-- Positive Losses: 143.9389
-- Dpo Losses: 0.6869
-- Rewards/chosen: -1.4291
-- Rewards/rejected: -2.0251
-- Rewards/accuracies: 0.6627
-- Rewards/margins: 0.5959
-- Rewards/margins Max: 2.2953
-- Rewards/margins Min: -1.1073
-- Rewards/margins Std: 1.5052
-- Logps/rejected: -461.6887
-- Logps/chosen: -428.1342
-- Logits/rejected: -2.2738
-- Logits/chosen: -2.3165
+- Loss: 14.3852
+- Positive Losses: 141.6597
+- Dpo Losses: 0.6849
+- Rewards/chosen: -1.4061
+- Rewards/rejected: -2.0012
+- Rewards/accuracies: 0.6667
+- Rewards/margins: 0.5951
+- Rewards/margins Max: 2.2885
+- Rewards/margins Min: -1.0995
+- Rewards/margins Std: 1.4978
+- Logps/rejected: -459.3069
+- Logps/chosen: -425.8328
+- Logits/rejected: -2.2783
+- Logits/chosen: -2.3207
 
 ## Model description
 

all_results.json

@@ -1,5 +1,23 @@
 {
     "epoch": 3.0,
+    "eval_dpo_losses": 0.6849049925804138,
+    "eval_logits/chosen": -2.3206958770751953,
+    "eval_logits/rejected": -2.278275489807129,
+    "eval_logps/chosen": -425.83282470703125,
+    "eval_logps/rejected": -459.3068542480469,
+    "eval_loss": 14.385190963745117,
+    "eval_positive_losses": 141.65965270996094,
+    "eval_rewards/accuracies": 0.6666666865348816,
+    "eval_rewards/chosen": -1.406116247177124,
+    "eval_rewards/margins": 0.5951257348060608,
+    "eval_rewards/margins_max": 2.288450241088867,
+    "eval_rewards/margins_min": -1.0994930267333984,
+    "eval_rewards/margins_std": 1.49783194065094,
+    "eval_rewards/rejected": -2.00124192237854,
+    "eval_runtime": 280.9524,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 7.119,
+    "eval_steps_per_second": 0.224,
     "train_loss": 0.36563416943303856,
     "train_runtime": 9271.2095,
     "train_samples": 5678,

config.json

@@ -0,0 +1,40 @@
+{
+  "_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,
+  "quantization_config": {
+    "_load_in_4bit": true,
+    "_load_in_8bit": false,
+    "bnb_4bit_compute_dtype": "bfloat16",
+    "bnb_4bit_quant_type": "nf4",
+    "bnb_4bit_use_double_quant": false,
+    "llm_int8_enable_fp32_cpu_offload": false,
+    "llm_int8_has_fp16_weight": false,
+    "llm_int8_skip_modules": null,
+    "llm_int8_threshold": 6.0,
+    "load_in_4bit": true,
+    "load_in_8bit": false,
+    "quant_method": "bitsandbytes"
+  },
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 10000.0,
+  "sliding_window": 4096,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.39.0.dev0",
+  "use_cache": true,
+  "vocab_size": 32000
+}

eval_results.json

@@ -0,0 +1,21 @@
+{
+    "epoch": 3.0,
+    "eval_dpo_losses": 0.6849049925804138,
+    "eval_logits/chosen": -2.3206958770751953,
+    "eval_logits/rejected": -2.278275489807129,
+    "eval_logps/chosen": -425.83282470703125,
+    "eval_logps/rejected": -459.3068542480469,
+    "eval_loss": 14.385190963745117,
+    "eval_positive_losses": 141.65965270996094,
+    "eval_rewards/accuracies": 0.6666666865348816,
+    "eval_rewards/chosen": -1.406116247177124,
+    "eval_rewards/margins": 0.5951257348060608,
+    "eval_rewards/margins_max": 2.288450241088867,
+    "eval_rewards/margins_min": -1.0994930267333984,
+    "eval_rewards/margins_std": 1.49783194065094,
+    "eval_rewards/rejected": -2.00124192237854,
+    "eval_runtime": 280.9524,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 7.119,
+    "eval_steps_per_second": 0.224
+}

runs/Jul29_11-59-46_node13/events.out.tfevents.1722232048.node13.208681.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27cf773fbb08166d927057ef8860f0f46f0e43436887295078d5f278b9524a14
+size 1130