End of training

README.md

@@ -3,9 +3,12 @@ library_name: peft
 license: apache-2.0
 base_model: mistralai/Mistral-7B-v0.1
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+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. -->
 
 # zephyr-7b-dpo-qlora
 
-This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on an unknown dataset.
+This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-qlora](https://huggingface.co/alignment-handbook/zephyr-7b-sft-qlora) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.5035
-- Rewards/chosen: -2.0213
-- Rewards/rejected: -3.0170
-- Rewards/accuracies: 0.7656
-- Rewards/margins: 0.9957
-- Logps/rejected: -549.2363
-- Logps/chosen: -448.5603
-- Logits/rejected: -1.1850
-- Logits/chosen: -1.2569
+- Loss: 0.5031
+- Rewards/chosen: -1.9728
+- Rewards/rejected: -2.9618
+- Rewards/accuracies: 0.7695
+- Rewards/margins: 0.9890
+- Logps/rejected: -543.7128
+- Logps/chosen: -443.7073
+- Logits/rejected: -1.1810
+- Logits/chosen: -1.2525
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 0.9984301412872841,
+    "eval_logits/chosen": -1.2525343894958496,
+    "eval_logits/rejected": -1.1810396909713745,
+    "eval_logps/chosen": -443.7073059082031,
+    "eval_logps/rejected": -543.7128295898438,
+    "eval_loss": 0.5031002759933472,
+    "eval_rewards/accuracies": 0.76953125,
+    "eval_rewards/chosen": -1.9727791547775269,
+    "eval_rewards/margins": 0.9890241622924805,
+    "eval_rewards/rejected": -2.961803436279297,
+    "eval_runtime": 110.64,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 18.077,
+    "eval_steps_per_second": 0.289,
     "total_flos": 0.0,
     "train_loss": 0.5421361258444796,
     "train_runtime": 7660.497,

config.json

@@ -0,0 +1,42 @@
+{
+  "_name_or_path": "mistralai/Mistral-7B-v0.1",
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "head_dim": 128,
+  "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_storage": "uint8",
+    "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": "4.45.2",
+  "use_cache": true,
+  "vocab_size": 32000
+}

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 0.9984301412872841,
+    "eval_logits/chosen": -1.2525343894958496,
+    "eval_logits/rejected": -1.1810396909713745,
+    "eval_logps/chosen": -443.7073059082031,
+    "eval_logps/rejected": -543.7128295898438,
+    "eval_loss": 0.5031002759933472,
+    "eval_rewards/accuracies": 0.76953125,
+    "eval_rewards/chosen": -1.9727791547775269,
+    "eval_rewards/margins": 0.9890241622924805,
+    "eval_rewards/rejected": -2.961803436279297,
+    "eval_runtime": 110.64,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 18.077,
+    "eval_steps_per_second": 0.289
+}

runs/Oct28_13-44-52_TUEILTI-GPU004/events.out.tfevents.1730127744.TUEILTI-GPU004.852476.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77d43d21111c23324268f0b21cde9dc6e4bcdad03cc1dd07c33c243d29de6876
+size 828