Spark-X2.6-1.7B, compiled for the browser

XHToken's Spark-X2.6-1.7B compiled into a fusor-ai bundle: the weights the browser runtime of spark-x2-6-wasm loads to hold a text conversation on the device's CPU. Try it in the Space fusorlabs/spark-x2-6-wasm.

Source: XHToken/Spark-X2.6-1.7B at revision f18c032c7675f807c0fdffe281dbe124edba2605, compiled with the q7 profile: every large matrix 7-bit (1921 MB).
Format: manifest.json lists every tensor and part; each part is named by its BLAKE3 hash, which the runtime checks as it loads.
Reproduce it: just model in spark-x2-6-wasm downloads the checkpoint and compiles the same bundle.

This is not a PyTorch or safetensors checkpoint: it only loads in the spark-x2-6-wasm runtime.

The model is Apache-2.0 (LICENSE).

Downloads last month: -; Downloads are not tracked for this model. How to track

Model tree for fusorlabs/spark-x2.6-1.7b-web

Base model

Finetuned

Finetuned

(4)

this model