|
--- |
|
tags: |
|
- gptq |
|
- 4bit |
|
- int4 |
|
- gptqmodel |
|
- modelcloud |
|
- llama-3.1 |
|
- 70b |
|
- instruct |
|
license: llama3.1 |
|
--- |
|
This model has been quantized using [GPTQModel](https://github.com/ModelCloud/GPTQModel). |
|
|
|
- **bits**: 4 |
|
- **group_size**: 128 |
|
- **desc_act**: true |
|
- **static_groups**: false |
|
- **sym**: true |
|
- **lm_head**: false |
|
- **damp_percent**: 0.0025 |
|
- **true_sequential**: true |
|
- **model_name_or_path**: "" |
|
- **model_file_base_name**: "model" |
|
- **quant_method**: "gptq" |
|
- **checkpoint_format**: "gptq" |
|
- **meta**： |
|
- **quantizer**: "gptqmodel:0.9.9-dev0" |
|
|
|
## Example: |
|
```python |
|
from transformers import AutoTokenizer |
|
from gptqmodel import GPTQModel |
|
|
|
model_name = "ModelCloud/Meta-Llama-3.1-70B-Instruct-gptq-4bit" |
|
|
|
prompt = [{"role": "user", "content": "I am in Shanghai, preparing to visit the natural history museum. Can you tell me the best way to"}] |
|
|
|
tokenizer = AutoTokenizer.from_pretrained(model_name) |
|
|
|
model = GPTQModel.from_quantized(model_name) |
|
|
|
input_tensor = tokenizer.apply_chat_template(prompt, add_generation_prompt=True, return_tensors="pt") |
|
outputs = model.generate(input_ids=input_tensor.to(model.device), max_new_tokens=100) |
|
result = tokenizer.decode(outputs[0][input_tensor.shape[1]:], skip_special_tokens=True) |
|
|
|
print(result) |
|
``` |
|
|
|
## lm-eval benchmark: |
|
|
|
``` |
|
| Tasks |Version|Filter|n-shot| Metric | |Value | |Stderr| |
|
|---------------------------------------|------:|------|-----:|----------|---|-----:|---|-----:| |
|
|arc_challenge | 1|none | 0|acc |↑ |0.6186|± |0.0142| |
|
| | |none | 0|acc_norm |↑ |0.6297|± |0.0141| |
|
|arc_easy | 1|none | 0|acc |↑ |0.8628|± |0.0071| |
|
| | |none | 0|acc_norm |↑ |0.8338|± |0.0076| |
|
|boolq | 2|none | 0|acc |↑ |0.8761|± |0.0058| |
|
|hellaswag | 1|none | 0|acc |↑ |0.6463|± |0.0048| |
|
| | |none | 0|acc_norm |↑ |0.8389|± |0.0037| |
|
|lambada_openai | 1|none | 0|acc |↑ |0.7561|± |0.0060| |
|
| | |none | 0|perplexity|↓ |3.0311|± |0.0639| |
|
|mmlu | 1|none | |acc |↑ |0.8100|± |0.0032| |
|
| - humanities | 1|none | |acc |↑ |0.7981|± |0.0057| |
|
| - formal_logic | 0|none | 0|acc |↑ |0.6349|± |0.0431| |
|
| - high_school_european_history | 0|none | 0|acc |↑ |0.8545|± |0.0275| |
|
| - high_school_us_history | 0|none | 0|acc |↑ |0.9412|± |0.0165| |
|
| - high_school_world_history | 0|none | 0|acc |↑ |0.9198|± |0.0177| |
|
| - international_law | 0|none | 0|acc |↑ |0.9008|± |0.0273| |
|
| - jurisprudence | 0|none | 0|acc |↑ |0.8796|± |0.0315| |
|
| - logical_fallacies | 0|none | 0|acc |↑ |0.8650|± |0.0268| |
|
| - moral_disputes | 0|none | 0|acc |↑ |0.8266|± |0.0204| |
|
| - moral_scenarios | 0|none | 0|acc |↑ |0.8559|± |0.0117| |
|
| - philosophy | 0|none | 0|acc |↑ |0.8360|± |0.0210| |
|
| - prehistory | 0|none | 0|acc |↑ |0.8827|± |0.0179| |
|
| - professional_law | 0|none | 0|acc |↑ |0.6675|± |0.0120| |
|
| - world_religions | 0|none | 0|acc |↑ |0.9181|± |0.0210| |
|
| - other | 1|none | |acc |↑ |0.8304|± |0.0064| |
|
| - business_ethics | 0|none | 0|acc |↑ |0.7900|± |0.0409| |
|
| - clinical_knowledge | 0|none | 0|acc |↑ |0.8566|± |0.0216| |
|
| - college_medicine | 0|none | 0|acc |↑ |0.7630|± |0.0324| |
|
| - global_facts | 0|none | 0|acc |↑ |0.5800|± |0.0496| |
|
| - human_aging | 0|none | 0|acc |↑ |0.8206|± |0.0257| |
|
| - management | 0|none | 0|acc |↑ |0.8835|± |0.0318| |
|
| - marketing | 0|none | 0|acc |↑ |0.9231|± |0.0175| |
|
| - medical_genetics | 0|none | 0|acc |↑ |0.9400|± |0.0239| |
|
| - miscellaneous | 0|none | 0|acc |↑ |0.9144|± |0.0100| |
|
| - nutrition | 0|none | 0|acc |↑ |0.8660|± |0.0195| |
|
| - professional_accounting | 0|none | 0|acc |↑ |0.6454|± |0.0285| |
|
| - professional_medicine | 0|none | 0|acc |↑ |0.8971|± |0.0185| |
|
| - virology | 0|none | 0|acc |↑ |0.5602|± |0.0386| |
|
| - social sciences | 1|none | |acc |↑ |0.8736|± |0.0059| |
|
| - econometrics | 0|none | 0|acc |↑ |0.7018|± |0.0430| |
|
| - high_school_geography | 0|none | 0|acc |↑ |0.9242|± |0.0189| |
|
| - high_school_government_and_politics| 0|none | 0|acc |↑ |0.9741|± |0.0115| |
|
| - high_school_macroeconomics | 0|none | 0|acc |↑ |0.8410|± |0.0185| |
|
| - high_school_microeconomics | 0|none | 0|acc |↑ |0.8992|± |0.0196| |
|
| - high_school_psychology | 0|none | 0|acc |↑ |0.9229|± |0.0114| |
|
| - human_sexuality | 0|none | 0|acc |↑ |0.8779|± |0.0287| |
|
| - professional_psychology | 0|none | 0|acc |↑ |0.8497|± |0.0145| |
|
| - public_relations | 0|none | 0|acc |↑ |0.7273|± |0.0427| |
|
| - security_studies | 0|none | 0|acc |↑ |0.8163|± |0.0248| |
|
| - sociology | 0|none | 0|acc |↑ |0.9154|± |0.0197| |
|
| - us_foreign_policy | 0|none | 0|acc |↑ |0.9300|± |0.0256| |
|
| - stem | 1|none | |acc |↑ |0.7456|± |0.0075| |
|
| - abstract_algebra | 0|none | 0|acc |↑ |0.6300|± |0.0485| |
|
| - anatomy | 0|none | 0|acc |↑ |0.7926|± |0.0350| |
|
| - astronomy | 0|none | 0|acc |↑ |0.8947|± |0.0250| |
|
| - college_biology | 0|none | 0|acc |↑ |0.9444|± |0.0192| |
|
| - college_chemistry | 0|none | 0|acc |↑ |0.5800|± |0.0496| |
|
| - college_computer_science | 0|none | 0|acc |↑ |0.6700|± |0.0473| |
|
| - college_mathematics | 0|none | 0|acc |↑ |0.5400|± |0.0501| |
|
| - college_physics | 0|none | 0|acc |↑ |0.6275|± |0.0481| |
|
| - computer_security | 0|none | 0|acc |↑ |0.8200|± |0.0386| |
|
| - conceptual_physics | 0|none | 0|acc |↑ |0.7830|± |0.0269| |
|
| - electrical_engineering | 0|none | 0|acc |↑ |0.7862|± |0.0342| |
|
| - elementary_mathematics | 0|none | 0|acc |↑ |0.7593|± |0.0220| |
|
| - high_school_biology | 0|none | 0|acc |↑ |0.9194|± |0.0155| |
|
| - high_school_chemistry | 0|none | 0|acc |↑ |0.7143|± |0.0318| |
|
| - high_school_computer_science | 0|none | 0|acc |↑ |0.9200|± |0.0273| |
|
| - high_school_mathematics | 0|none | 0|acc |↑ |0.5185|± |0.0305| |
|
| - high_school_physics | 0|none | 0|acc |↑ |0.6556|± |0.0388| |
|
| - high_school_statistics | 0|none | 0|acc |↑ |0.7361|± |0.0301| |
|
| - machine_learning | 0|none | 0|acc |↑ |0.7054|± |0.0433| |
|
|openbookqa | 1|none | 0|acc |↑ |0.3660|± |0.0216| |
|
| | |none | 0|acc_norm |↑ |0.4620|± |0.0223| |
|
|piqa | 1|none | 0|acc |↑ |0.8264|± |0.0088| |
|
| | |none | 0|acc_norm |↑ |0.8319|± |0.0087| |
|
|rte | 1|none | 0|acc |↑ |0.7184|± |0.0271| |
|
|truthfulqa_mc1 | 2|none | 0|acc |↑ |0.3917|± |0.0171| |
|
|winogrande | 1|none | 0|acc |↑ |0.7924|± |0.0114| |
|
|
|
| Groups |Version|Filter|n-shot|Metric| |Value | |Stderr| |
|
|------------------|------:|------|------|------|---|-----:|---|-----:| |
|
|mmlu | 1|none | |acc |↑ |0.8100|± |0.0032| |
|
| - humanities | 1|none | |acc |↑ |0.7981|± |0.0057| |
|
| - other | 1|none | |acc |↑ |0.8304|± |0.0064| |
|
| - social sciences| 1|none | |acc |↑ |0.8736|± |0.0059| |
|
| - stem | 1|none | |acc |↑ |0.7456|± |0.0075| |
|
``` |
