Instructions to use tsavage68/Error_Q1.5_1000steps_1e6rate_SFT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tsavage68/Error_Q1.5_1000steps_1e6rate_SFT with Transformers:

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

pipe = pipeline("text-generation", model="tsavage68/Error_Q1.5_1000steps_1e6rate_SFT")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("tsavage68/Error_Q1.5_1000steps_1e6rate_SFT")
model = AutoModelForCausalLM.from_pretrained("tsavage68/Error_Q1.5_1000steps_1e6rate_SFT")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use tsavage68/Error_Q1.5_1000steps_1e6rate_SFT with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tsavage68/Error_Q1.5_1000steps_1e6rate_SFT"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tsavage68/Error_Q1.5_1000steps_1e6rate_SFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/tsavage68/Error_Q1.5_1000steps_1e6rate_SFT

SGLang

How to use tsavage68/Error_Q1.5_1000steps_1e6rate_SFT 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 "tsavage68/Error_Q1.5_1000steps_1e6rate_SFT" \
    --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": "tsavage68/Error_Q1.5_1000steps_1e6rate_SFT",
		"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 "tsavage68/Error_Q1.5_1000steps_1e6rate_SFT" \
        --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": "tsavage68/Error_Q1.5_1000steps_1e6rate_SFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use tsavage68/Error_Q1.5_1000steps_1e6rate_SFT with Docker Model Runner:
```
docker model run hf.co/tsavage68/Error_Q1.5_1000steps_1e6rate_SFT
```

Error_Q1.5_1000steps_1e6rate_SFT

This model is a fine-tuned version of deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 1.7311

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: 1e-06
train_batch_size: 2
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 4
optimizer: Use adafactor and the args are: No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 100
training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss
3.258	0.4	25	3.3120
3.2475	0.8	50	3.2579
3.1416	1.192	75	3.1255
2.9461	1.592	100	2.9168
2.7616	1.992	125	2.6859
2.498	2.384	150	2.4867
2.3356	2.784	175	2.3327
2.219	3.176	200	2.2049
2.1203	3.576	225	2.1078
2.0088	3.976	250	2.0359
1.9985	4.368	275	1.9716
1.9447	4.768	300	1.9176
1.836	5.16	325	1.8782
1.8332	5.5600	350	1.8498
1.7718	5.96	375	1.8258
1.7685	6.352	400	1.8005
1.7063	6.752	425	1.7825
1.7479	7.144	450	1.7717
1.6852	7.5440	475	1.7629
1.7561	7.944	500	1.7562
1.7028	8.336	525	1.7506
1.6847	8.736	550	1.7465
1.7223	9.128	575	1.7423
1.7248	9.528	600	1.7400
1.6413	9.928	625	1.7374
1.694	10.32	650	1.7359
1.6753	10.72	675	1.7343
1.6786	11.112	700	1.7333
1.6643	11.512	725	1.7330
1.6458	11.912	750	1.7324
1.6171	12.304	775	1.7315
1.6835	12.704	800	1.7313
1.6772	13.096	825	1.7313
1.6967	13.496	850	1.7312
1.6791	13.896	875	1.7314
1.6434	14.288	900	1.7312
1.6684	14.688	925	1.7314
1.6245	15.08	950	1.7312
1.6731	15.48	975	1.7311
1.7103	15.88	1000	1.7311

Framework versions

Transformers 4.49.0
Pytorch 2.6.0+cu124
Datasets 3.3.1
Tokenizers 0.21.0

Downloads last month: 4

Safetensors

Model size

2B params

Tensor type

F16

Model tree for tsavage68/Error_Q1.5_1000steps_1e6rate_SFT

Base model

deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B

Finetuned

(642)

this model