Instructions to use kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000 with Transformers:

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

pipe = pipeline("text-generation", model="kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000")
model = AutoModelForCausalLM.from_pretrained("kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000

SGLang

How to use kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000 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 "kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000" \
    --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": "kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000",
		"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 "kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000" \
        --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": "kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000 with Docker Model Runner:
```
docker model run hf.co/kashif/Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000
```

Qwen3-0.6B-h128-l4-a16-ctx256-pred64-vocab4096-MeanScaleUniform-lr1.0e-05-bs16-steps1000 / training_config.json

kashif HF Staff

Upload 4 files

074b52e verified about 1 year ago

Raw

History Blame Contribute Delete

1.06 kB

	{
	"pretrained_model_name_or_path": "Qwen/Qwen3-0.6B",
	"output_dir": "./outputs",
	"seed": 42,
	"tf32": true,
	"model_type": "causal",
	"vocab_size": 4096,
	"hidden_size": 256,
	"num_hidden_layers": 8,
	"num_attention_heads": 8,
	"context_length": 256,
	"prediction_length": 64,
	"tokenizer_class": "NonUniformBins",
	"binning_power": 2.0,
	"exponential_base": 1.01,
	"n_special_tokens": 2,
	"pad_token_id": 0,
	"eos_token_id": 1,
	"use_eos_token": true,
	"min_past": 64,
	"drop_prob": 0.1,
	"shuffle_buffer_length": 20000,
	"per_device_train_batch_size": 32,
	"learning_rate": 0.0001,
	"max_steps": 40000,
	"warmup_ratio": 0.1,
	"lr_scheduler_type": "cosine",
	"optim": "adamw_torch",
	"gradient_accumulation_steps": 4,
	"log_steps": 20,
	"save_steps": 100,
	"dataloader_num_workers": 8,
	"torch_compile": true,
	"logger": "<Logger __main__ (INFO)>",
	"total_train_batch_size": 128,
	"short_model_name": "Qwen3-0.6B",
	"run_name": "Qwen3-0.6B-h256-l8-a8-ctx256-pred64-vocab4096-NonUniform-lr1.0e-04-bs128-steps40000"
	}