Instructions to use willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine with Transformers:

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

pipe = pipeline("text-generation", model="willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine")
model = AutoModelForCausalLM.from_pretrained("willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine", device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine

SGLang

How to use willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine 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 "willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine" \
    --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": "willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine",
		"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 "willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine" \
        --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": "willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine with Docker Model Runner:
```
docker model run hf.co/willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine
```

Qwen3-8B-SDFT-Math-LoRA-pristine / README.md

willamazon1

Merged math GRPO-LoRA (iter_0000120) into the SFT base

63d4d62 verified about 1 month ago

preview code

Raw History Blame Contribute Delete

2.62 kB

---
license: apache-2.0
base_model: Qwen/Qwen3-8B-Base
library_name: transformers
pipeline_tag: text-generation
tags:
- math
- reasoning
- reinforcement-learning
- lora
- grpo
---
# Qwen3-8B SDFT + math GRPO-LoRA (rollout 120)

Qwen3-8B (dense, 36 layers, 8.19B params). **Fully merged bf16 weights** — the LoRA adapter is
folded into the base matrices, so this is a drop-in replacement for `Qwen/Qwen3-8B-Base`; no PEFT
adapter loading required.

## Training pipeline

1. **Base**: `Qwen/Qwen3-8B-Base`.
2. **SFT**: supervised fine-tuning on a balanced oracle mixture (math / retrieval-augmented QA /
   tool-use dialogue).
3. **Math RL** (this checkpoint): GRPO on math problems with a verifiable boxed-answer reward,
   training LoRA adapters (rank 128, alpha 128 ⇒ scaling 1.0) on `q/k/v`, `o_proj`,
   `gate/up_proj` and `down_proj` of every layer, for 120 rollouts. DAPO-style recipe
   (clip-higher 0.2/0.28, dynamic sampling, token-level policy-gradient loss) with truncated
   importance sampling for the train/rollout mismatch.

Trained with [slime](https://github.com/THUDM/slime) on Megatron-LM (TP=4, CP=1, bf16).

## Honest evaluation note

Over these 120 rollouts the RL stage was **stable but did not measurably improve** the
training-set pass@1 (0.558 → 0.578, linear slope +4e-5 per rollout against a per-rollout standard
deviation of 0.027 — the change is inside the noise). Prompts solved 16/16 rose slightly (50.6 →
56.9 of 256 sampled) and prompts never solved were flat. Treat this as **the SFT model plus a small,
non-degrading RL delta**, not as a demonstrated improvement over the SFT base. It is published for
reproducibility and comparison, and it is the first checkpoint of this line whose frozen base
weights are provably uncorrupted.

## Usage

```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(repo, dtype=torch.bfloat16).to("cuda")
prompt = "Question: What is 12*8?\nAnswer:"
ids = tok(prompt, return_tensors="pt").input_ids.cuda()
print(tok.decode(model.generate(ids, max_new_tokens=64, do_sample=False)[0][ids.shape[1]:],
                 skip_special_tokens=True))
```

This is a **base-style (completion)** model, not an instruction-tuned chat model: prompt it with
`Question: … \nAnswer:` style completions rather than a chat template.

## Limitations

Inherits the biases and knowledge cutoff of `Qwen3-8B-Base`. Math answers are not guaranteed
correct — verify before relying on them.

	---
	license: apache-2.0
	base_model: Qwen/Qwen3-8B-Base
	library_name: transformers
	pipeline_tag: text-generation
	tags:
	- math
	- reasoning
	- reinforcement-learning
	- lora
	- grpo
	---

	# Qwen3-8B SDFT + math GRPO-LoRA (rollout 120)

	Qwen3-8B (dense, 36 layers, 8.19B params). Fully merged bf16 weights — the LoRA adapter is
	folded into the base matrices, so this is a drop-in replacement for `Qwen/Qwen3-8B-Base`; no PEFT
	adapter loading required.

	## Training pipeline

	1. Base: `Qwen/Qwen3-8B-Base`.
	2. SFT: supervised fine-tuning on a balanced oracle mixture (math / retrieval-augmented QA /
	tool-use dialogue).
	3. Math RL (this checkpoint): GRPO on math problems with a verifiable boxed-answer reward,
	training LoRA adapters (rank 128, alpha 128 ⇒ scaling 1.0) on `q/k/v`, `o_proj`,
	`gate/up_proj` and `down_proj` of every layer, for 120 rollouts. DAPO-style recipe
	(clip-higher 0.2/0.28, dynamic sampling, token-level policy-gradient loss) with truncated
	importance sampling for the train/rollout mismatch.

	Trained with [slime](https://github.com/THUDM/slime) on Megatron-LM (TP=4, CP=1, bf16).

	## Honest evaluation note

	Over these 120 rollouts the RL stage was stable but did not measurably improve the
	training-set pass@1 (0.558 → 0.578, linear slope +4e-5 per rollout against a per-rollout standard
	deviation of 0.027 — the change is inside the noise). Prompts solved 16/16 rose slightly (50.6 →
	56.9 of 256 sampled) and prompts never solved were flat. Treat this as **the SFT model plus a small,
	non-degrading RL delta**, not as a demonstrated improvement over the SFT base. It is published for
	reproducibility and comparison, and it is the first checkpoint of this line whose frozen base
	weights are provably uncorrupted.

	## Usage

	```python
	import torch
	from transformers import AutoModelForCausalLM, AutoTokenizer

	repo = "willamazon1/Qwen3-8B-SDFT-Math-LoRA-pristine"
	tok = AutoTokenizer.from_pretrained(repo)
	model = AutoModelForCausalLM.from_pretrained(repo, dtype=torch.bfloat16).to("cuda")

	prompt = "Question: What is 12*8?\nAnswer:"
	ids = tok(prompt, return_tensors="pt").input_ids.cuda()
	print(tok.decode(model.generate(ids, max_new_tokens=64, do_sample=False)[0][ids.shape[1]:],
	skip_special_tokens=True))
	```

	This is a base-style (completion) model, not an instruction-tuned chat model: prompt it with
	`Question: … \nAnswer:` style completions rather than a chat template.

	## Limitations

	Inherits the biases and knowledge cutoff of `Qwen3-8B-Base`. Math answers are not guaranteed
	correct — verify before relying on them.