tahamajs/RR / Docs /module_metadata /flat_llm.json
tahamajs's picture
download
raw
607 Bytes
{
"module_name": "flat_llm",
"class_name": "FLATLLMCompressor",
"category": "compression",
"family": "compression",
"params": [
"model",
"tokenizer",
"keep_ratio",
"damping",
"calibration_loader",
"calibration_texts",
"device",
"structured_24"
],
"ops": [
"linear",
"matmul",
"norm",
"softmax"
],
"docstring": "FLAT-LLM Compressor: Activation-aware neuron and output channel pruning.",
"rank": 16,
"d_state": 1,
"num_experts": 1,
"top_k": 1,
"sparsity_pct": 0.0,
"bits": 4,
"hash_val": 252752936575530323565886621594541529646
}

Xet Storage Details

Size:
607 Bytes
·
Xet hash:
e2504c8a652c4c2774b1f17ceebaa22db5d8e8d895e44c18dd5aa64374c3038e

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.