Instructions to use tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000")
model = AutoModelForCausalLM.from_pretrained("tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000

SGLang

How to use tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000 with Docker Model Runner:
```
docker model run hf.co/tiendung/vietmistral_enbagel_lora-ctx4096-cpt3000
```

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

data enbagel.jsonl.zst, 160k english instructs, 378M

{'loss': 0.6286, 'learning_rate': 5.076582709326072e-05, 'epoch': 1.0}
{'loss': 0.6117, 'learning_rate': 5.074880871341049e-05, 'epoch': 1.0}
{'loss': 0.4023, 'learning_rate': 5.073179033356025e-05, 'epoch': 1.0}
{'loss': 0.7861, 'learning_rate': 5.071477195371e-05, 'epoch': 1.0}
{'loss': 0.5591, 'learning_rate': 5.069775357385977e-05, 'epoch': 1.0}
{'loss': 0.5298, 'learning_rate': 5.0680735194009535e-05, 'epoch': 1.0}
{'loss': 0.7576, 'learning_rate': 5.066371681415929e-05, 'epoch': 1.0}
{'loss': 0.6725, 'learning_rate': 5.064669843430906e-05, 'epoch': 1.0}
{'loss': 0.4667, 'learning_rate': 5.062968005445882e-05, 'epoch': 1.0}
 51%|█████     | 3000/5936 [12:06:13<11:35:20, 14.21s/it]finetune_qlora_vietmistra.sh: line 33: 91840 Killed

Downloads last month: 18