tim-lawson's picture
Push model using huggingface_hub.
2a39252 verified
raw
history blame contribute delete
150 Bytes
{
"auxk": 256,
"dead_steps_threshold": 76,
"dead_threshold": 0.001,
"k": 32,
"n_inputs": 512,
"n_latents": 32768,
"standardize": false
}