Qwen3-8B-Maxima_A1
Qwen3-8B fine-tuned for tool calling and agentic workflows on a 5,000-sample subset of the SFT split of Toucan-1.5M.
Given a set of tool definitions, the model emits <tool_call> JSON, chains calls across steps using tool results, and answers directly when no tool is relevant. It is a small-scale proof of concept, released as merged 16-bit weights.
Training
| Base model | Qwen3-8B |
| Method | QLoRA (4-bit), rank 16, alpha 16, all attention and MLP projections; merged to 16-bit |
| Data | 5,000 samples from Toucan-1.5M (SFT split): 2,000 multi-turn, 1,250 single-turn (original), 750 single-turn (diversified), 1,000 irrelevant-tool (no call expected) |
| Sequence length | Up to 4,096 tokens |
| Schedule | 1 epoch (625 steps), learning rate 2e-4 (cosine), effective batch size 8 |
| Loss | Assistant responses only |
| Hardware | 1 × NVIDIA T4, Unsloth |
Usage
The 16-bit weights are about 15 GB. On smaller GPUs, load the model in 4-bit (for example with bitsandbytes).
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "Rumiii/Qwen3-8B-Maxima_A1"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(repo, dtype="auto", device_map="auto")
tools = [{
"type": "function",
"function": {
"name": "get_weather",
"description": "Get the current weather for a city.",
"parameters": {
"type": "object",
"properties": {"city": {"type": "string"}},
"required": ["city"],
},
},
}]
messages = [{"role": "user", "content": "What's the weather in Hyderabad?"}]
prompt = tok.apply_chat_template(messages, tools=tools, tokenize=False, add_generation_prompt=True)
inputs = tok(prompt, return_tensors="pt").to(model.device)
out = model.generate(**inputs, max_new_tokens=256, do_sample=False)
print(tok.decode(out[0][inputs.input_ids.shape[1]:], skip_special_tokens=False).split("<|im_end|>")[0])
Expected output:
<tool_call>
{"name": "get_weather", "arguments": {"city": "Hyderabad"}}
</tool_call>
To continue after running the tool, append the call and its result to messages, rebuild the prompt with apply_chat_template, and generate again:
messages += [
{"role": "assistant", "content": "", "tool_calls": [
{"type": "function", "function": {"name": "get_weather", "arguments": {"city": "Hyderabad"}}}]},
{"role": "tool", "content": '{"temp_c": 31, "condition": "Partly cloudy"}'},
]
Limitations
- Trained on 5,000 samples only and not evaluated on public benchmarks.
- The training data is synthetic, generated from community MCP servers between June and September 2025.
- Always validate tool names and arguments before executing a tool call.
Acknowledgements
Data: Toucan-1.5M (Apache 2.0). Base model: Qwen3-8B (Apache 2.0).