|
--- |
|
language: |
|
- en |
|
license: creativeml-openrail-m |
|
library_name: transformers |
|
tags: |
|
- qlora |
|
- peft |
|
- prompts |
|
datasets: |
|
- knkarthick/dialogsum |
|
metrics: |
|
- accuracy |
|
base_model: TinyPixel/Llama-2-7B-bf16-sharded |
|
--- |
|
## Training procedure |
|
|
|
|
|
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 |
|
### Framework versions |
|
|
|
|
|
- PEFT 0.4.0 |
|
|
|
``` |
|
# adding back the LoRA adopters to the base Llama-2 model |
|
|
|
lora_config = LoraConfig.from_pretrained('Andyrasika/qlora-dialogue-summary') |
|
model = get_peft_model(model, lora_config) |
|
|
|
inputs = tokenizer(text, return_tensors="pt") |
|
outputs = model.generate(input_ids=inputs['input_ids'], attention_mask=inputs['attention_mask'], max_new_tokens=100 ,repetition_penalty=1.2) |
|
print(tokenizer.decode(outputs[0], skip_special_tokens=True)) |
|
|
|
``` |