End of training

README.md

@@ -1,8 +1,12 @@
 ---
 base_model: HuggingFaceH4/zephyr-7b-gemma-sft-v0.1
+datasets:
+- argilla/dpo-mix-7k
+- RedaAlami/PKU-SafeRLHF-Processed
 library_name: peft
 license: other
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
@@ -16,17 +20,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-gemma-dpo
 
-This model is a fine-tuned version of [HuggingFaceH4/zephyr-7b-gemma-sft-v0.1](https://huggingface.co/HuggingFaceH4/zephyr-7b-gemma-sft-v0.1) on the None dataset.
+This model is a fine-tuned version of [HuggingFaceH4/zephyr-7b-gemma-sft-v0.1](https://huggingface.co/HuggingFaceH4/zephyr-7b-gemma-sft-v0.1) on the argilla/dpo-mix-7k and the RedaAlami/PKU-SafeRLHF-Processed datasets.
 It achieves the following results on the evaluation set:
-- Loss: 0.6449
-- Rewards/chosen: -0.3403
-- Rewards/rejected: -0.5854
-- Rewards/accuracies: 0.6065
-- Rewards/margins: 0.2451
-- Logps/rejected: -334.6864
-- Logps/chosen: -295.8665
-- Logits/rejected: 436.0117
-- Logits/chosen: 452.6395
+- Loss: 0.6478
+- Rewards/chosen: -0.3452
+- Rewards/rejected: -0.5788
+- Rewards/accuracies: 0.6169
+- Rewards/margins: 0.2336
+- Logps/rejected: -334.5554
+- Logps/chosen: -295.9647
+- Logits/rejected: 436.0139
+- Logits/chosen: 452.6414
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.9986987638256344,
+    "eval_logits/chosen": 452.641357421875,
+    "eval_logits/rejected": 436.01385498046875,
+    "eval_logps/chosen": -295.9647216796875,
+    "eval_logps/rejected": -334.555419921875,
+    "eval_loss": 0.6477869153022766,
+    "eval_rewards/accuracies": 0.6168639063835144,
+    "eval_rewards/chosen": -0.34524381160736084,
+    "eval_rewards/margins": 0.23360556364059448,
+    "eval_rewards/rejected": -0.5788493752479553,
+    "eval_runtime": 243.7753,
+    "eval_samples": 5406,
+    "eval_samples_per_second": 22.176,
+    "eval_steps_per_second": 0.693,
     "total_flos": 0.0,
     "train_loss": 0.6547494133313497,
     "train_runtime": 8032.2795,

config.json

@@ -0,0 +1,44 @@
+{
+  "_name_or_path": "HuggingFaceH4/zephyr-7b-gemma-sft-v0.1",
+  "architectures": [
+    "GemmaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 2,
+  "eos_token_id": 1,
+  "head_dim": 256,
+  "hidden_act": "gelu",
+  "hidden_activation": "gelu_pytorch_tanh",
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 24576,
+  "max_position_embeddings": 8192,
+  "model_type": "gemma",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 16,
+  "pad_token_id": 0,
+  "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-06,
+  "rope_scaling": null,
+  "rope_theta": 10000.0,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.43.3",
+  "use_cache": true,
+  "vocab_size": 256000
+}

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 1.9986987638256344,
+    "eval_logits/chosen": 452.641357421875,
+    "eval_logits/rejected": 436.01385498046875,
+    "eval_logps/chosen": -295.9647216796875,
+    "eval_logps/rejected": -334.555419921875,
+    "eval_loss": 0.6477869153022766,
+    "eval_rewards/accuracies": 0.6168639063835144,
+    "eval_rewards/chosen": -0.34524381160736084,
+    "eval_rewards/margins": 0.23360556364059448,
+    "eval_rewards/rejected": -0.5788493752479553,
+    "eval_runtime": 243.7753,
+    "eval_samples": 5406,
+    "eval_samples_per_second": 22.176,
+    "eval_steps_per_second": 0.693
+}

runs/Jul31_11-58-27_ip-172-16-2-184.us-west-2.compute.internal/events.out.tfevents.1722435568.ip-172-16-2-184.us-west-2.compute.internal.45409.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43428576dd032014b432311e40581bf451bca7298c6ac252912e58f8d45e1d44
+size 828