aufklarer's picture
Add Core ML INT8 model
2957d4b verified
Raw
History Blame Contribute Delete
1.26 kB
{
"audio_token_id": 151671,
"audio_tokens_per_second": 12.5,
"cache_update": "scatter_along_axis",
"context_note": "The fixed 1024-token MLState cache covers at most about 1.4 minutes of raw audio before prompt/output tokens.",
"decoder_split": {
"enabled": false,
"part1_layers": 28,
"part2_layers": 0,
"split_at": null
},
"enumerated_t": [
1,
128
],
"files": {
"model": "decoder.mlmodelc"
},
"fused_embedding": false,
"head_dim": 128,
"hidden_size": 1024,
"intermediate_size": 3072,
"io_precision": "float16",
"max_seq_length": 1024,
"model_type": "moss-transcribe-diarize-decoder-coreml",
"multifunction": {
"decoder_function": "decoder",
"embedding_function": "embedding",
"file": "decoder.mlmodelc",
"shared_weight": "embed_tokens.weight/lm_head.weight"
},
"num_heads": 16,
"num_kv_heads": 8,
"num_layers": 28,
"quantization": "linear-int8-block32",
"recommended_optimization_hints": {
"specializationStrategy": "FastPrediction"
},
"rms_norm_eps": 1e-06,
"rope_theta": 1000000.0,
"shape_mode": "range",
"source_model": "OpenMOSS-Team/MOSS-Transcribe-Diarize",
"source_revision": "e6d68cdfcddbdad1a7e8454f0cb859cad76e2502",
"vocab_size": 151936
}