ramithuh/evodiff-webgpu

AI model by ramithuh

Open original source ↗

From the publisher

Model card & documentation

Source preview

Read the publisher’s intended use, setup instructions, evaluations and limitations. The original model card is the source of truth.

A complete, standalone port of OADM38M: token embedding, input projection, all 16 ByteNet residual blocks, final LayerNorm, 31-way decoder, and order-agnostic iterative sampling. Runtime files are plain ES modules with no npm dependencies, Node built-ins, WASM, Python, or external inference runtime. All neural operations in the WebGPU engine execute WGSL compute shaders. The separate CPU implementation is a reference backend. The source of truth is the pinned 38M config, checkpoint constructor, ByteNetTime, and ByteNetBlock / MaskedConv1d. The sampler follows the pinned unconditional and conditional OA-DM path: visit each masked position once in random order, rerun the entire network, softmax the selected position, and unmask it. Upstream uses len(MSAAAS)-6 = 21 categories, including B; this port preserves that behavior. Fixed motif positions remain unchanged. Replay replaces framework-specific multinomial RNG with recorded inverse-CDF draws and FP64 JS/Python softmax; native PyTorch RNG seeds are not interchangeable with the JS seed. Serve this directory…

ramithuh/evodiff-webgpu on Hugging Face A short preview, not the full document.

Read the full model card ↗ · Preview checked 2026-11-22T08:44:13.238Z

Inside the original model card — Document outline
  1. EvoDiff-Seq 38M in WebGPU
  2. Provenance and architecture audit
  3. Quick start
  4. API contract
  5. Reproduce the export and oracles
  6. Measured validation
  7. Export sizes and quantization error versus original FP32
  8. Mode Matrix: worst final-output relRMS
  9. NVIDIA GB10 latency
  10. WGSL portability
  11. Hardware browser invocation
  12. Vendor sync and denovo.run mapping

Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.

Documentation belongs to its respective authors. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.

Model overview

ramithuh/evodiff-webgpu is a task-unspecified model repository published by ramithuh on Hugging Face. A library was not reported.

This page summarizes Hub metadata. For intended use, training data, evaluation results and limitations, consult the original model card.

Model facts

Task
Not reported
Library
Not reported
Recent downloads (30 days)
0
Cumulative likes
0
Hugging Face trending score
0
Reported safetensors parameters
Not reported
Architecture
Not reported
License
Not reported — check the model card
Access
Not gated by Hugging Face
Created
2026-11-22T06:54:08.000Z
Last modified
2026-11-22T06:54:14.000Z

Compatibility and lineage

Language coverage was not reported.

Base-model lineage was not included.

Parameter count is not a RAM/VRAM requirement. Check precision, quantization, context length and runtime compatibility in the model card; no hardware or API-cost claim is inferred here.

Use and evaluate

Check license terms and access requirements first. Review the model card’s documented loading instructions, evaluations and safety limitations. Benchmark scores are not imported by this directory, and popularity is not an accuracy ranking.

No model weights are downloaded or executed by AltAPIs. Never enable remote model code without reviewing it.

Model card and usage instructions

Model files and configuration

Community discussion

Source and freshness

Source: Hugging Face Hub. Metadata observed 2026-11-22T07:03:33.420Z. Daily imports are snapshots, not real-time monitoring.

Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.

Open original source

Related models