Instructions to use w-ahmad/tiny-llama-baseline-gelu with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use w-ahmad/tiny-llama-baseline-gelu with Transformers:

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

pipe = pipeline("text-generation", model="w-ahmad/tiny-llama-baseline-gelu")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("w-ahmad/tiny-llama-baseline-gelu", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use w-ahmad/tiny-llama-baseline-gelu with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "w-ahmad/tiny-llama-baseline-gelu"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "w-ahmad/tiny-llama-baseline-gelu",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/w-ahmad/tiny-llama-baseline-gelu

SGLang

How to use w-ahmad/tiny-llama-baseline-gelu 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 "w-ahmad/tiny-llama-baseline-gelu" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "w-ahmad/tiny-llama-baseline-gelu",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "w-ahmad/tiny-llama-baseline-gelu" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "w-ahmad/tiny-llama-baseline-gelu",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use w-ahmad/tiny-llama-baseline-gelu with Docker Model Runner:
```
docker model run hf.co/w-ahmad/tiny-llama-baseline-gelu
```

tiny-llama-baseline-gelu

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 1.3872

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: 0.0003
train_batch_size: 32
eval_batch_size: 32
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 64
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: constant
training_steps: 10000

Training results

Training Loss	Epoch	Step	Validation Loss
5.8115	0.0159	250	2.8801
4.6262	0.0319	500	2.3344
4.1792	0.0478	750	2.0890
3.8695	0.0637	1000	1.9427
3.7163	0.0796	1250	1.8483
3.5153	0.0956	1500	1.7793
3.4517	0.1115	1750	1.7311
3.3712	0.1274	2000	1.6902
3.3071	0.1433	2250	1.6564
3.1952	0.1593	2500	1.6281
3.1764	0.1752	2750	1.6048
3.1841	0.1911	3000	1.5861
3.1887	0.2071	3250	1.5709
3.1254	0.2230	3500	1.5566
3.0764	0.2389	3750	1.5394
3.0250	0.2548	4000	1.5282
3.0381	0.2708	4250	1.5174
2.9927	0.2867	4500	1.5064
2.9577	0.3026	4750	1.4978
2.9401	0.3185	5000	1.4891
2.9238	0.3345	5250	1.4809
2.9697	0.3504	5500	1.4734
2.9711	0.3663	5750	1.4655
2.9360	0.3823	6000	1.4597
2.9155	0.3982	6250	1.4540
2.8983	0.4141	6500	1.4485
2.8937	0.4300	6750	1.4428
2.8779	0.4460	7000	1.4374
2.8592	0.4619	7250	1.4312
2.8556	0.4778	7500	1.4266
2.8630	0.4937	7750	1.4223
2.7939	0.5097	8000	1.4177
2.8342	0.5256	8250	1.4145
2.8549	0.5415	8500	1.4098
2.8061	0.5574	8750	1.4052
2.8270	0.5734	9000	1.4023
2.7998	0.5893	9250	1.3985
2.7635	0.6052	9500	1.3957
2.7969	0.6212	9750	1.3913
2.7639	0.6371	10000	1.3872

Framework versions

Transformers 5.15.0.dev0
Pytorch 2.6.0+cu124
Datasets 5.0.1
Tokenizers 0.22.2

Downloads last month: 18

Safetensors

Model size

34.2M params

Tensor type

F32