Instructions to use devorein/llama_7b-instruct_lora_int4-subj_eval with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use devorein/llama_7b-instruct_lora_int4-subj_eval with Transformers:

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

pipe = pipeline("text-generation", model="devorein/llama_7b-instruct_lora_int4-subj_eval")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("devorein/llama_7b-instruct_lora_int4-subj_eval")
model = AutoModelForCausalLM.from_pretrained("devorein/llama_7b-instruct_lora_int4-subj_eval")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use devorein/llama_7b-instruct_lora_int4-subj_eval with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "devorein/llama_7b-instruct_lora_int4-subj_eval"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "devorein/llama_7b-instruct_lora_int4-subj_eval",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/devorein/llama_7b-instruct_lora_int4-subj_eval

SGLang

How to use devorein/llama_7b-instruct_lora_int4-subj_eval 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 "devorein/llama_7b-instruct_lora_int4-subj_eval" \
    --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": "devorein/llama_7b-instruct_lora_int4-subj_eval",
		"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 "devorein/llama_7b-instruct_lora_int4-subj_eval" \
        --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": "devorein/llama_7b-instruct_lora_int4-subj_eval",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use devorein/llama_7b-instruct_lora_int4-subj_eval with Docker Model Runner:
```
docker model run hf.co/devorein/llama_7b-instruct_lora_int4-subj_eval
```

llama_7b-instruct_lora_int4-subj_eval / xturing.json

devorein

Upload folder using huggingface_hub

5a34da1 about 3 years ago

raw

history blame

673 Bytes

	{
	"model_name": "llama_lora_int4",
	"finetuning_config": {
	"learning_rate": 0.0001,
	"gradient_accumulation_steps": 1,
	"batch_size": 16,
	"weight_decay": 0.01,
	"warmup_steps": 50,
	"eval_steps": 5000,
	"save_steps": 5000,
	"max_length": 256,
	"num_train_epochs": 10,
	"logging_steps": 10,
	"max_grad_norm": 2.0,
	"save_total_limit": 4,
	"optimizer_name": "adamw",
	"output_dir": "saved_model"
	},
	"generation_config": {
	"penalty_alpha": 0.6,
	"top_k": 4,
	"max_new_tokens": 256,
	"do_sample": false,
	"top_p": null
	}
	}