alkahest-0.8b

Browser-oriented ONNX export of a Qwen3.5 Heretic checkpoint packaged for WebGPU / Transformers.js.

Capabilities

  • Supported inputs: text, image
  • Not included in this package: video

Version Notes

  • This is the stable v1 browser package.
  • For the enhanced multimodal variant, see thomasjvu/alkahest-0.8b-v2.
  • The v2 package adds video for this model family.

Provenance

  • Source model: tvall43/Qwen3.5-0.8B-heretic-v3
  • Base model for inherited processor assets: Qwen/Qwen3.5-0.8B
  • Architecture family: qwen3_5_conditional_generation
  • Expected architecture: Qwen3_5ForConditionalGeneration
  • Target dtype: q4f16
  • Target device: webgpu

Expected ONNX Sessions

  • vision_encoder_q4f16.onnx
  • embed_tokens_q4f16.onnx
  • decoder_model_merged_q4f16.onnx

Usage

Load this repo with Transformers.js using the model's transformers.js_config metadata and WebGPU backend.

Downloads last month
23
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for thomasjvu/alkahest-0.8b

Quantized
(4)
this model