|
--- |
|
base_model: Qwen/Qwen2.5-7B-Instruct |
|
library_name: peft |
|
license: apache-2.0 |
|
datasets: |
|
- medalpaca/medical_meadow_medical_flashcards |
|
language: |
|
- en |
|
pipeline_tag: text-generation |
|
--- |
|
|
|
# Model Card for FlowerTune-Qwen2.5-7B-Instruct-Medical-PEFT |
|
|
|
This PEFT adapter has been trained by using [Flower](https://flower.ai/), a friendly federated AI framework. |
|
|
|
The adapter and benchmark results have been submitted to the [FlowerTune LLM Medical Leaderboard](https://flower.ai/benchmarks/llm-leaderboard/medical/). |
|
|
|
|
|
## Model Details |
|
|
|
Please check the following GitHub project for model details and evaluation results: |
|
|
|
[https://github.com/mrs83/FlowerTune-Qwen2.5-7B-Instruct-Medical](https://github.com/mrs83/FlowerTune-Qwen2.5-7B-Instruct-Medical) |
|
|
|
|
|
## Training procedure |
|
|
|
|
|
The following `bitsandbytes` quantization config was used during training: |
|
- quant_method: bitsandbytes |
|
- _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: fp4 |
|
- bnb_4bit_use_double_quant: False |
|
- bnb_4bit_compute_dtype: float32 |
|
- bnb_4bit_quant_storage: uint8 |
|
- load_in_4bit: True |
|
- load_in_8bit: False |
|
|
|
### Framework versions |
|
|
|
|
|
- PEFT 0.6.2 |
|
- Flower 1.12.0 |
