ananas0

Hakam-0.8b-sportsqa-distill-lora

README

License: apache-2.0

Try it now

Live demo on a free Colab T4 (opens a Gradio app with match replays pre-loaded):

Or load it yourself:

python
import torch
from transformers import AutoModelForImageTextToText, AutoProcessor
from peft import PeftModel

base = "Qwen/Qwen3.5-0.8B"
processor = AutoProcessor.from_pretrained(base, trust_remote_code=True)
model = AutoModelForImageTextToText.from_pretrained(
    base, dtype=torch.bfloat16, device_map="cuda", trust_remote_code=True)
model = PeftModel.from_pretrained(model, "ananas0/Hakam-0.8b-sportsqa-distill-lora")
model.eval()

Video input: 2 fps, max 32 frames, max_pixels=256*28*28. The model's sweet spot is clips of 8–16 seconds with one action in them.

The numbers

Measured on 1,000 held-out Sports-QA test questions (fuzzy match against gold labels, threshold 0.80):

Table with columns: model, params, accuracy
model	params	accuracy
Qwen3.5 0.8B, plain	0.8B	22.7%
0.8B + this adapter	0.8B	39.9%
Qwen3.5 27B (the teacher)	27B	39.7%

The student and teacher are statistically tied (n=1000, ±3pp 95% CI). On counterfactual questions the student scores higher than the teacher (82% vs 70%).

How it was trained

Recipe: free-generation distillation. The 27B teacher answered ~11,700 training questions with a fixed analyst-style prompt; its per-token top-32 logits were recorded. The student was then teacher-forced on those generations with loss = 0.7·KL + 0.3·CE at Hinton temperature 2.0.
Adapter: LoRA r=16, α=32, dropout 0.05 on all attention + FFN projections of the language model. Vision encoder frozen. 6.4M trainable parameters (0.74%).
Training: 3 epochs, 35,016 steps, AdamW lr=1e-4, batch size 1, bf16.
Data: a balanced subset of Sports-QA: 8 sports (football, basketball, volleyball, aerobic gymnastics, vault, uneven bars, balance beam, floor exercise), 200 training videos per sport.

Full pipeline, evaluation code and figures: github.com/ANANAS-HA/HAKAM

Demo replay clips: ananas0/hakam-demo-clips

Limitations

Domain-specific: trained only on Sports-QA footage and question styles. It will describe other video reasonably (that's the base model) but the distillation gain is sports-specific.
Temporal/causal questions whose gold answers use FineGym taxonomy terms remain hard; the teacher itself doesn't know that vocabulary, and a student can't learn what its teacher doesn't know.
Clips longer than ~16 seconds get sampled sparsely (32 frames at 2 fps) and answer quality degrades.

Credits

Graduation project. Built on Qwen3.5 (Alibaba) and the Sports-QA dataset. Apache 2.0.

Available on FriendliAI

Dedicated Endpoints

Run this model inference on single tenant GPU with unmatched speed and reliability at scale.

Learn more

Model Details

Model Provider

ananas0

Model Tree

Base

Qwen/Qwen3.5-0.8B

Adapter

this model

Input Modalities

TextImageVideo

Output Modalities

Text

Supported Functionality