CHIH-HUNG/llama-2-13b-FINETUNE1_17w-r16

Model Card for Model ID

在llama-2-13b上使用huangyt/FINETUNE1資料集進行訓練，總資料筆數約17w

Fine-Tuning Information

• GPU: RTX4090 (single core / 24564MiB)
• model: meta-llama/Llama-2-13b-hf
• dataset: huangyt/FINETUNE1 (共約17w筆訓練集)
• peft_type: LoRA
• lora_rank: 16
• lora_target: gate_proj, up_proj, down_proj
• per_device_train_batch_size: 8
• gradient_accumulation_steps: 8
• learning_rate : 5e-5
• epoch: 1
• precision: bf16
• quantization: load_in_4bit

Fine-Tuning Detail

• train_loss: 0.66
• train_runtime: 16:26:58 (use deepspeed)

Evaluation

• 評估結果來自HuggingFaceH4/open_llm_leaderboard
• 與Llama-2-13b比較4種Benchmark，包含ARC、HellaSwag、MMLU、TruthfulQA

Model	Average	ARC	HellaSwag	MMLU	TruthfulQA
meta-llama/Llama-2-13b-hf	56.9	58.11	80.97	54.34	34.17
meta-llama/Llama-2-13b-chat-hf	59.93	59.04	81.94	54.64	44.12
CHIH-HUNG/llama-2-13b-Fintune_1_17w	58.24	59.47	81	54.31	38.17
CHIH-HUNG/llama-2-13b-huangyt_Fintune_1_17w-q_k_v_o_proj	58.49	59.73	81.06	54.53	38.64
CHIH-HUNG/llama-2-13b-Fintune_1_17w-gate_up_down_proj	58.81	57.17	82.26	55.89	39.93
CHIH-HUNG/llama-2-13b-FINETUNE1_17w-r16	58.86	57.25	82.27	56.16	39.75
CHIH-HUNG/llama-2-13b-FINETUNE1_17w-r4	58.71	56.74	82.27	56.18	39.65

How to convert dataset to json

• 在load_dataset中輸入資料集名稱，並且在take中輸入要取前幾筆資料
• 觀察該資料集的欄位名稱，填入example欄位中(例如system_prompt、question、response)
• 最後指定json檔儲存位置 (json_filename)

import json
from datasets import load_dataset

# 讀取數據集，take可以取得該數據集前n筆資料
dataset = load_dataset("huangyt/FINETUNE1", split="train", streaming=True)

# 提取所需欄位並建立新的字典列表
extracted_data = []
for example in dataset:
    extracted_example = {
        "instruction": example["instruction"],
        "input": example["input"],
        "output": example["output"]
    }
    extracted_data.append(extracted_example)

# 指定 JSON 文件名稱
json_filename = "huangyt_FINETUNE_1.json"

# 寫入 JSON 文件
with open(json_filename, "w") as json_file:
    json.dump(extracted_data, json_file, indent=4)

print(f"數據已提取並保存為 {json_filename}")