Instructions to use onuralp/Llama-2-13b-QLoRA-02 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use onuralp/Llama-2-13b-QLoRA-02 with Transformers:

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

pipe = pipeline("text-generation", model="onuralp/Llama-2-13b-QLoRA-02")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("onuralp/Llama-2-13b-QLoRA-02")
model = AutoModelForCausalLM.from_pretrained("onuralp/Llama-2-13b-QLoRA-02", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use onuralp/Llama-2-13b-QLoRA-02 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "onuralp/Llama-2-13b-QLoRA-02"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "onuralp/Llama-2-13b-QLoRA-02",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/onuralp/Llama-2-13b-QLoRA-02

SGLang

How to use onuralp/Llama-2-13b-QLoRA-02 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 "onuralp/Llama-2-13b-QLoRA-02" \
    --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": "onuralp/Llama-2-13b-QLoRA-02",
		"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 "onuralp/Llama-2-13b-QLoRA-02" \
        --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": "onuralp/Llama-2-13b-QLoRA-02",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use onuralp/Llama-2-13b-QLoRA-02 with Docker Model Runner:
```
docker model run hf.co/onuralp/Llama-2-13b-QLoRA-02
```

qlora-out

This model is a fine-tuned version of meta-llama/Llama-2-13b-hf on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.5407

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 0.0004
train_batch_size: 2
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 8
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 300
num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
0.8973	0.03	20	0.7029
0.6828	0.06	40	0.6521
0.6521	0.09	60	0.6199
0.7857	0.11	80	0.6066
0.6208	0.14	100	0.6063
0.6805	0.17	120	0.5969
0.5928	0.2	140	0.5989
0.715	0.23	160	0.5844
0.5647	0.26	180	0.5979
0.6778	0.29	200	0.5889
0.5907	0.31	220	0.5772
0.5536	0.34	240	0.5917
0.7422	0.37	260	0.6781
0.6328	0.4	280	0.5785
0.5705	0.43	300	0.5720
0.6124	0.46	320	0.5753
0.4735	0.49	340	0.6203
0.4602	0.52	360	0.5772
0.8571	0.54	380	0.5750
0.5504	0.57	400	0.6040
0.6307	0.6	420	0.5796
0.4782	0.63	440	0.5639
0.4159	0.66	460	0.5689
0.6393	0.69	480	0.5661
0.8243	0.72	500	0.5698
0.4744	0.74	520	0.5536
0.4395	0.77	540	0.5536
0.543	0.8	560	0.5493
0.4451	0.83	580	0.5421
0.5384	0.86	600	0.5467
0.4438	0.89	620	0.5379
0.4168	0.92	640	0.5398
0.469	0.94	660	0.5402
0.6766	0.97	680	0.5407

Framework versions

Transformers 4.34.1
Pytorch 2.0.1+cu118
Datasets 2.14.6
Tokenizers 0.14.1

Downloads last month: 8

Model tree for onuralp/Llama-2-13b-QLoRA-02

Base model

meta-llama/Llama-2-13b-hf

Quantized

(31)

this model

Collection including onuralp/Llama-2-13b-QLoRA-02

neurips-2023-llm-efficiency

Collection

Fine-tune models, datasets and artifacts used for llm efficiency competition. https://llm-efficiency-challenge.github.io/challenge • 15 items • Updated Oct 27, 2023