steven0226

qwen3vl-8b-chartqa-merged-16bit

README

License: apache-2.0

Training summary

15,000 shuffled ChartQA training examples
1 epoch; maximum sequence length 2,048
LoRA rank 16, alpha 16, dropout 0
Vision, language, attention, and MLP modules enabled for LoRA adaptation
Effective A100 batch size 16
8-bit AdamW; peak learning rate 2e-4; linear schedule
Runtime 3,579 seconds; final training loss 0.5907

Evaluation

Complete ChartQA test set, isolated vLLM evaluation, relaxed accuracy with 5% numeric tolerance:

Table with columns: Split, n, Accuracy
Split	n	Accuracy
Human	1,250	77.28%
Augmented	1,250	95.20%
Overall	2,500	86.24%

Evaluation revision: 519060ef43df3261e0512e5ae4c82a4d4e675f32.

The downstream AWQ checkpoint scored 85.52%, a -0.72 pp change that passed the predefined -2 pp quality gate.

Serving reference

The formal A100 benchmark used this checkpoint as the 16-bit baseline. Weight files total 17.53 GB.

Table with columns: Concurrency, Output tok/s, TTFT p95, TPOT p95, E2E p95
Concurrency	Output tok/s	TTFT p95	TPOT p95	E2E p95
1	67.29	160.76 ms	13.43 ms/tok	1,007.16 ms
4	231.02	299.78 ms	15.04 ms/tok	1,169.61 ms
8	387.58	473.68 ms	18.33 ms/tok	1,426.90 ms

Complete comparison and reproducibility artifacts are in the AWQ repo bench/.

Usage

This is a standard merged Qwen3-VL checkpoint. It can be loaded with a Qwen3-VL-compatible Transformers/vLLM environment. The verified serving stack was vLLM 0.25.1+cu129 with torch 2.11.0+cu129 on Linux A100.

python
from vllm import LLM

llm = LLM(
    model="steven0226/qwen3vl-8b-chartqa-merged-16bit",
    revision="519060ef43df3261e0512e5ae4c82a4d4e675f32",
    max_model_len=4096,
    limit_mm_per_prompt={"image": 1, "video": 0},
)

For a complete image/chat example, use the AWQ model card and replace the model ID/revision.

Limitations

Evaluated primarily on English ChartQA short answers.
Not validated for general OCR documents, out-of-domain charts, or safety-critical use.
This checkpoint is the larger quality reference; the AWQ checkpoint is recommended for serving when its -0.72 pp quality tradeoff is acceptable.

中文摘要

這是將 ChartQA LoRA adapter 合併回 Qwen3-VL-8B 後的 16-bit 完整模型，也是 AWQ 量化前的品質與 serving baseline。完整 2,500 題 vLLM 評估為 86.24%；權重檔 17.53 GB。若部署可接受 0.72 個百分點的下降，建議改用 7.55 GB 的 AWQ 版本。

Available on FriendliAI

Dedicated Endpoints

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

Learn more

Container

Run this model inference with full control and performance in your environment.

Learn more

Model Details

Model Provider

steven0226

Model Tree

Base

unsloth/Qwen3-VL-8B-Instruct-unsloth-bnb-4bit

Fine-tuned

this model

Input Modalities

TextImage

Output Modalities

Text

Supported Functionality

Dedicated EndpointsContainer

Explore FriendliAI today

Get started Talk to an engineer