Downloads · 30 days
7
29% of all-time downloads
CodeLlama-7b-Test-Instruct-lora
zzzghttt/CodeLlama-7b-Test-Instruct-lora
CodeLlama-7b-Test-Instruct-lora is a machine learning model from zzzghttt. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for peft.
This repo contains a low-rank adapter for CodeLlama-7b-Instruct fit on the zzzghttt/code2test dataset. The Lora model is primarily aimed at generating high-quality unit tests in Java.
All-time downloads
24
Public
Repo size
403 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pt269 MB · 67%
At a glance
- Library
- peft
- Access
- Public
- Created
- Dec 30, 2023
- Updated
- Feb 3, 2024
- SHA
- 3e5f5ca0
Model card
From the Hugging Face model README
CodeLlama-7b-Test-Instruct-lora
Description
This repo contains a low-rank adapter for CodeLlama-7b-Instruct fit on the zzzghttt/code2test dataset. The Lora model is primarily aimed at generating high-quality unit tests in Java.
How to use
Training data
Training procedure
This version of the weights was trained with the following hyperparameters:
- batch_size: 128
- micro_batch_size: 4
- num_epochs: 3 (load from best epoch)
- learning_rate: 3e-4
- cutoff_len: 2048
- lora_r: 64
- lora_alpha: 16
- lora_dropout: 0.05
- lora_target_modules: ['q_proj', 'v_proj']
The following bitsandbytes quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: True
- load_in_4bit: False
- 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: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
Framework versions
- PEFT 0.5.0
- Library
- peft
- Created
- Dec 30, 2023
- Updated
- Feb 3, 2024