Instructions to use d4rkninja/tanpo-product with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use d4rkninja/tanpo-product with Transformers:

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

pipe = pipeline("text-generation", model="d4rkninja/tanpo-product")
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("d4rkninja/tanpo-product")
model = AutoModelForCausalLM.from_pretrained("d4rkninja/tanpo-product", 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 d4rkninja/tanpo-product with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "d4rkninja/tanpo-product"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "d4rkninja/tanpo-product",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/d4rkninja/tanpo-product

SGLang

How to use d4rkninja/tanpo-product 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 "d4rkninja/tanpo-product" \
    --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": "d4rkninja/tanpo-product",
		"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 "d4rkninja/tanpo-product" \
        --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": "d4rkninja/tanpo-product",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use d4rkninja/tanpo-product with Docker Model Runner:
```
docker model run hf.co/d4rkninja/tanpo-product
```

d4rkninja commited on 26 days ago

Commit

ce66046

verified ·

1 Parent(s): 9d96cd6

Upload evaluation/COMPARE_BASE.md with huggingface_hub

Browse files

Files changed (1) hide show

evaluation/COMPARE_BASE.md +45 -12

evaluation/COMPARE_BASE.md CHANGED Viewed

@@ -1,12 +1,45 @@
-# DarkLab Base vs Fine-tune Comparison
-Official DarkLab 20-task domain rubric comparison against `unsloth/LFM2.5-1.2B-Instruct`.
-| Domain | Base rubric | Fine-tune rubric | Delta vs base |
-|---|---:|---:|---:|
-| **Product — format-v1** | 92.6% | **99.4% (99.38%)** | **+6.8 pts** |
-| Hiring | 83.8% | **92.9%** | **+9.2 pts** |
-Product format-v1 quality: **7.98** vs base **7.35**. Hiring row retained from the existing comparison: **92.9% vs 83.8%**.
-Automated rubric results are directional; they do not replace human judgment.

+# Base vs Fine-tune comparison
+**Generated:** 2026-09-17T22:14:33+05:30
+**Base model:** `unsloth/LFM2.5-1.2B-Instruct`
+**Same tasks + same scorer as domain PEFT suite**
+## Overall
+| Domain | Base LFM2.5 | Fine-tune | Delta |
+|---|---:|---:|---:|
+| tanpo-product | 92.6% | **88.4%** | -4.2 (trails) |
+| tanpo-hiring | 83.8% | **92.9%** | +9.2 (beats) |
+| Domain | Base quality | FT quality |
+|---|---:|---:|
+| tanpo-product | 7.35 | 6.84 |
+| tanpo-hiring | 6.40 | 7.18 |
+## tanpo-product per-category
+| Category | Base % | FT % | Delta |
+|---|---:|---:|---:|
+| `competitive_positioning` | 93.8 | 87.5 | -6.2 |
+| `crisis_tradeoffs` | 100.0 | 91.7 | -8.3 |
+| `customer_discovery` | 87.5 | 83.3 | -4.2 |
+| `gtm_product_loops` | 93.8 | 81.2 | -12.5 |
+| `metrics_okrs` | 100.0 | 88.9 | -11.1 |
+| `org_operating_cadence` | 93.8 | 88.2 | -5.5 |
+| `pricing_monetization` | 88.9 | 88.9 | +0.0 |
+| `product_improvement` | 83.3 | 87.5 | +4.2 |
+| `roadmap_sequencing` | 93.8 | 100.0 | +6.2 |
+## tanpo-hiring per-category
+| Category | Base % | FT % | Delta |
+|---|---:|---:|---:|
+| `debrief_calibration` | 83.3 | 87.5 | +4.2 |
+| `hiring_scorecards` | 81.2 | 100.0 | +18.8 |
+| `interview_loop_design` | 85.2 | 100.0 | +14.8 |
+| `job_descriptions` | 88.2 | 100.0 | +11.8 |
+| `offer_framing` | 81.2 | 87.5 | +6.2 |
+| `outbound_recruiting` | 93.8 | 100.0 | +6.2 |
+| `rejection_pipeline_ops` | 62.5 | 87.5 | +25.0 |
+| `screening_scripts` | 81.2 | 100.0 | +18.8 |
+| `structured_interview_kits` | 85.2 | 77.8 | -7.4 |