End of training

README.md

@@ -1,7 +1,10 @@
 ---
 base_model: TII-Frontier-Team/falcon3-3b-instruct
+datasets:
+- TII-Frontier-Team/Reasoning_DPO
 library_name: peft
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
@@ -15,17 +18,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-dpo-qlora
 
-This model is a fine-tuned version of [TII-Frontier-Team/falcon3-3b-instruct](https://huggingface.co/TII-Frontier-Team/falcon3-3b-instruct) on an unknown dataset.
+This model is a fine-tuned version of [TII-Frontier-Team/PEFT-falcon3b-it-gsm8k](https://huggingface.co/TII-Frontier-Team/PEFT-falcon3b-it-gsm8k) on the TII-Frontier-Team/Reasoning_DPO dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0287
-- Rewards/chosen: -4.6985
-- Rewards/rejected: -10.6531
-- Rewards/accuracies: 0.9276
-- Rewards/margins: 5.9547
-- Logps/rejected: -1101.2122
-- Logps/chosen: -502.6163
-- Logits/rejected: 1.9469
-- Logits/chosen: 2.1464
+- Loss: 0.0286
+- Rewards/chosen: -4.7078
+- Rewards/rejected: -10.6652
+- Rewards/accuracies: 0.9254
+- Rewards/margins: 5.9575
+- Logps/rejected: -1102.4209
+- Logps/chosen: -503.5470
+- Logits/rejected: 1.9412
+- Logits/chosen: 2.1408
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": 2.140789270401001,
+    "eval_logits/rejected": 1.9412086009979248,
+    "eval_logps/chosen": -503.5469665527344,
+    "eval_logps/rejected": -1102.4208984375,
+    "eval_loss": 0.028598472476005554,
+    "eval_rewards/accuracies": 0.9253731369972229,
+    "eval_rewards/chosen": -4.707759380340576,
+    "eval_rewards/margins": 5.957462787628174,
+    "eval_rewards/rejected": -10.66522216796875,
+    "eval_runtime": 214.7372,
+    "eval_samples": 21417,
+    "eval_samples_per_second": 99.736,
+    "eval_steps_per_second": 1.56,
     "total_flos": 0.0,
     "train_loss": 0.19036180805619987,
     "train_runtime": 15997.0818,

config.json

@@ -0,0 +1,49 @@
+{
+  "_name_or_path": "TII-Frontier-Team/falcon3-3b-instruct",
+  "activation": "swiglu",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 10,
+  "eos_token_id": 127,
+  "head_dim": 256,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 9216,
+  "max_position_embeddings": 4096,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 22,
+  "num_key_value_heads": 4,
+  "output_hidden_states": true,
+  "pad_token_id": 11,
+  "parallel_attn": false,
+  "pretraining_tp": 1,
+  "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": 1000042,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.45.1",
+  "use_cache": true,
+  "vocab_size": 131072
+}

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 1.0,
+    "eval_logits/chosen": 2.140789270401001,
+    "eval_logits/rejected": 1.9412086009979248,
+    "eval_logps/chosen": -503.5469665527344,
+    "eval_logps/rejected": -1102.4208984375,
+    "eval_loss": 0.028598472476005554,
+    "eval_rewards/accuracies": 0.9253731369972229,
+    "eval_rewards/chosen": -4.707759380340576,
+    "eval_rewards/margins": 5.957462787628174,
+    "eval_rewards/rejected": -10.66522216796875,
+    "eval_runtime": 214.7372,
+    "eval_samples": 21417,
+    "eval_samples_per_second": 99.736,
+    "eval_steps_per_second": 1.56
+}

runs/Oct04_08-33-02_ip-172-16-43-187.us-west-2.compute.internal/events.out.tfevents.1728049146.ip-172-16-43-187.us-west-2.compute.internal.82163.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43d34f62343f56e86ff857383fb70db53f2cce9b0be6ecaf76afa1b191efebf4
+size 828