File size: 3,977 Bytes
43d3864 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 |
---
tags:
- generated_from_trainer
- code
- coding
- llama-2
model-index:
- name: Llama-2-7b-4bit-python-coder
results: []
license: apache-2.0
language:
- code
datasets:
- iamtarun/python_code_instructions_18k_alpaca
pipeline_tag: text-generation
---
# LlaMa 2 7b 4-bit Python Coder 👩💻
**LlaMa-2 7b** fine-tuned on the **CodeAlpaca 20k instructions dataset** by using the method **QLoRA** in 4-bit with [PEFT](https://github.com/huggingface/peft) library.
## Pretrained description
[Llama-2](https://huggingface.co/meta-llama/Llama-2-7b)
Meta developed and publicly released the Llama 2 family of large language models (LLMs), a collection of pretrained and fine-tuned generative text models ranging in scale from 7 billion to 70 billion parameters.
Model Architecture Llama 2 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align to human preferences for helpfulness and safety
## Training data
[python_code_instructions_18k_alpaca](https://huggingface.co/datasets/iamtarun/python_code_instructions_18k_alpaca)
The dataset contains problem descriptions and code in python language. This dataset is taken from sahil2801/code_instructions_120k, which adds a prompt column in alpaca style.
### Training hyperparameters
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float16
**SFTTrainer arguments**
```py
# Number of training epochs
num_train_epochs = 1
# Enable fp16/bf16 training (set bf16 to True with an A100)
fp16 = False
bf16 = True
# Batch size per GPU for training
per_device_train_batch_size = 4
# Number of update steps to accumulate the gradients for
gradient_accumulation_steps = 1
# Enable gradient checkpointing
gradient_checkpointing = True
# Maximum gradient normal (gradient clipping)
max_grad_norm = 0.3
# Initial learning rate (AdamW optimizer)
learning_rate = 2e-4
# Weight decay to apply to all layers except bias/LayerNorm weights
weight_decay = 0.001
# Optimizer to use
optim = "paged_adamw_32bit"
# Learning rate schedule
lr_scheduler_type = "cosine" #"constant"
# Ratio of steps for a linear warmup (from 0 to learning rate)
warmup_ratio = 0.03
```
### Framework versions
- PEFT 0.4.0
### Example of usage
```py
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
model_id = "mrm8488/llama-2-coder-7b"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id).to("cuda")
sample = dataset[randrange(len(dataset))]
prompt = f"""### Instruction:
Use the Task below and the Input given to write the Response, which is a programming code that can solve the Task.
### Task:
{sample['instruction']}
### Input:
{sample['input']}
### Response:
"""
input_ids = tokenizer(prompt, return_tensors="pt", truncation=True).input_ids.cuda()
# with torch.inference_mode():
outputs = model.generate(input_ids=input_ids, max_new_tokens=100, do_sample=True, top_p=0.9,temperature=0.5)
print(f"Prompt:\n{prompt}\n")
print(f"Generated instruction:\n{tokenizer.batch_decode(outputs.detach().cpu().numpy(), skip_special_tokens=True)[0][len(prompt):]}")
print(f"Ground truth:\n{sample['output']}")
```
### Citation
```
@misc {edumunozsala_2023,
author = { {Eduardo Muñoz} },
title = { llama-2-7b-int4-python-coder (Revision d30d193) },
year = 2023,
url = { https://huggingface.co/edumunozsala/llama-2-7b-int4-python-18k-alpaca },
doi = { 10.57967/hf/0931 },
publisher = { Hugging Face }
}
``` |