Instructions to use aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft with Transformers:

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

pipe = pipeline("text-generation", model="aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft")
model = AutoModelForCausalLM.from_pretrained("aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft")
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]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft

SGLang

How to use aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft 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 "aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft" \
    --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": "aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft",
		"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 "aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft" \
        --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": "aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft with Docker Model Runner:
```
docker model run hf.co/aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft
```

aadityabuilds commited on 4 days ago

Commit

8f19b4b

verified ·

1 Parent(s): 238a8e5

Update model card with KernelBook post-training description

Browse files

Files changed (1) hide show

README.md +52 -55

README.md CHANGED Viewed

@@ -1,75 +1,72 @@
 ---
 library_name: transformers
 tags:
-- generated_from_trainer
-model-index:
-- name: qwen2-5-coder-7b-kernelbook-sdft
-  results: []
 ---
-<!-- This model card has been generated automatically according to the information the Trainer had access to. You
-should probably proofread and complete it, then remove this comment. -->
-# qwen2-5-coder-7b-kernelbook-sdft
-This model was trained from scratch on the None dataset.
-It achieves the following results on the evaluation set:
-- Loss: 0.0272
-## 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: 5e-06
-- train_batch_size: 1
-- eval_batch_size: 1
-- seed: 42
-- distributed_type: multi-GPU
-- num_devices: 4
-- gradient_accumulation_steps: 2
-- total_train_batch_size: 8
-- total_eval_batch_size: 4
-- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
-- lr_scheduler_type: cosine
-- lr_scheduler_warmup_steps: 0.03
-- num_epochs: 1.0
-### Training results
-| Training Loss | Epoch  | Step | Validation Loss |
-|:-------------:|:------:|:----:|:---------------:|
-| 0.0646        | 0.0756 | 100  | 0.0782          |
-| 0.0499        | 0.1512 | 200  | 0.0564          |
-| 0.0459        | 0.2268 | 300  | 0.0496          |
-| 0.0402        | 0.3025 | 400  | 0.0421          |
-| 0.0386        | 0.3781 | 500  | 0.0394          |
-| 0.0417        | 0.4537 | 600  | 0.0357          |
-| 0.0290        | 0.5293 | 700  | 0.0332          |
-| 0.0266        | 0.6049 | 800  | 0.0310          |
-| 0.0242        | 0.6805 | 900  | 0.0298          |
-| 0.0194        | 0.7561 | 1000 | 0.0287          |
-| 0.0222        | 0.8318 | 1100 | 0.0277          |
-| 0.0213        | 0.9074 | 1200 | 0.0273          |
-| 0.0255        | 0.9830 | 1300 | 0.0272          |
-| 0.0190        | 1.0    | 1323 | 0.0272          |
-### Framework versions
-- Transformers 5.9.0
-- Pytorch 2.11.0+cu128
-- Datasets 4.8.5
-- Tokenizers 0.22.2

 ---
 library_name: transformers
+license: apache-2.0
+base_model: Qwen/Qwen2.5-Coder-7B-Instruct
 tags:
+- triton
+- kernelbook
+- code-generation
+- self-distillation
+- sdft
+- text-generation
+datasets:
+- custom
+language:
+- en
+pipeline_tag: text-generation
 ---
+# Qwen2.5-Coder-7B KernelBook SDFT
+**Self-Distillation Fine-Tuning (SDFT)** checkpoint of [Qwen/Qwen2.5-Coder-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct), post-trained on the **KernelBook** Triton kernel dataset.
+## Method
+This model was trained with **SDFT** (self-distillation fine-tuning): the student sees the user prompt plus privileged reference context (the target Triton implementation) and learns to reproduce the reference completion via forced-completion distillation (cross-entropy + KL on completion tokens). Training used a custom `KernelBookSDFTTrainer` on top of `transformers.Trainer` with DeepSpeed ZeRO-3.
+## Dataset
+- **KernelBook** — PyTorch module prompts paired with reference Triton kernels
+- Deduplicated, filtered to completions ≤4096 tokens, repo-stratified 80/10/10 split
+- **1 training epoch** on the KernelBook train split
+## Intended use
+Generate Triton GPU kernels from PyTorch-style module descriptions. Best for KernelBook-style conversion prompts; not evaluated as a general-purpose chat or reasoning model.
+## Quick start
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+model_id = "aadityabuilds/qwen2-5-coder-7b-kernelbook-sdft"
+tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
+model = AutoModelForCausalLM.from_pretrained(
+    model_id, torch_dtype="auto", device_map="auto", trust_remote_code=True
+)
+messages = [
+    {
+        "role": "user",
+        "content": "Convert the following PyTorch code to an equivalent Triton kernel...",
+    }
+]
+prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+outputs = model.generate(**inputs, max_new_tokens=1200, do_sample=False)
+print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[1] :], skip_special_tokens=True))
+```
+## Training summary
+| Setting | Value |
+|---------|-------|
+| Base model | Qwen2.5-Coder-7B-Instruct |
+| Method | SDFT (forced-completion distillation) |
+| Epochs | 1 |
+| Hardware | 4× H100 (Modal) |
+| Parallelism | DeepSpeed ZeRO-3, bf16 |
+## Limitations
+Specialized for KernelBook Triton codegen. May show reduced performance on general coding, math, and knowledge benchmarks compared to the base instruct model.