voice_audio_multi/model_specs/gigaam_asr.json
Fabio c36510c21e
Some checks failed
Linux Build / Linux cpu (push) Has been cancelled
Linux Build / Linux cpu-clang (push) Has been cancelled
Linux Build / Linux vulkan (push) Has been cancelled
mac-build / macOS CPU (push) Has been cancelled
Nix Build / Nix (cpu) (push) Has been cancelled
Nix Build / Nix (vulkan) (push) Has been cancelled
windows-build / Windows CPU (push) Has been cancelled
Initial import: audio.cpp with Orange Pi documentation
2026-11-04 19:05:22 +01:00

221 lines
5.2 KiB
JSON

{
"schema_version": 1,
"family": "gigaam_asr",
"display_name": "GigaAM ASR",
"description": "GigaAM v3 and Multilingual Conformer speech recognition with CTC or RNN-T decoding.",
"category": "asr",
"status": "supported",
"tasks": [
"asr"
],
"modes": [
"offline"
],
"languages": [
"ru",
"kk",
"ky",
"uz",
"en"
],
"capabilities": {
"asr": [
"word_timestamps",
"vad_chunking"
]
},
"dependencies": [],
"options": {
"request": [
{
"name": "return_timestamps",
"type": "bool",
"description": "Return word timestamps derived from decoder token emission frames.",
"required": false,
"default": false
},
{
"name": "audio_chunk_mode",
"type": "enum",
"description": "Long-form audio chunking mode.",
"values": [
"auto",
"fixed",
"quiet_energy",
"vad",
"none"
],
"required": false,
"default": "auto"
},
{
"name": "audio_chunk_duration_sec",
"type": "float",
"description": "Maximum duration of each audio chunk in seconds.",
"required": false,
"min": 0.02,
"max": 25,
"default": 25
}
],
"session": [
{
"name": "vad_model_path",
"type": "path",
"description": "Offline VAD model used only with audio_chunk_mode=vad.",
"required": false,
"default": ""
},
{
"name": "weight_type",
"type": "enum",
"description": "Weight storage type.",
"preset": "weight_type_full",
"required": false,
"default": "native"
}
],
"load": []
},
"runtime": {
"tags": [
"gguf"
]
},
"ui": {
"recommended_package": "gigaam_v3_ctc_f16",
"tags": [
"ASR",
"GGUF"
],
"docs": [
"docs/models/gigaam_asr.md"
]
},
"package_defaults": {
"download": {
"kind": "huggingface_snapshot",
"repo": "audio-cpp/audio.cpp-gguf",
"revision": "main",
"gated": false
}
},
"packages": [
{
"id": "gigaam_v3_ctc_f16",
"display_name": "GigaAM v3 CTC F16 GGUF",
"default": true,
"format": "gguf",
"precision": "f16",
"target_directory": "GigaAM-ASR-GGUF",
"files": [
"GigaAM-ASR-GGUF/gigaam-v3-ctc-f16.gguf"
],
"strip_prefix": "GigaAM-ASR-GGUF"
},
{
"id": "gigaam_v3_rnnt_f16",
"display_name": "GigaAM v3 RNN-T F16 GGUF",
"format": "gguf",
"precision": "f16",
"target_directory": "GigaAM-ASR-GGUF",
"files": [
"GigaAM-ASR-GGUF/gigaam-v3-rnnt-f16.gguf"
],
"strip_prefix": "GigaAM-ASR-GGUF"
},
{
"id": "gigaam_v3_e2e_ctc_f16",
"display_name": "GigaAM v3 E2E CTC F16 GGUF",
"format": "gguf",
"precision": "f16",
"target_directory": "GigaAM-ASR-GGUF",
"files": [
"GigaAM-ASR-GGUF/gigaam-v3-e2e-ctc-f16.gguf"
],
"strip_prefix": "GigaAM-ASR-GGUF"
},
{
"id": "gigaam_v3_e2e_rnnt_f16",
"display_name": "GigaAM v3 E2E RNN-T F16 GGUF",
"format": "gguf",
"precision": "f16",
"target_directory": "GigaAM-ASR-GGUF",
"files": [
"GigaAM-ASR-GGUF/gigaam-v3-e2e-rnnt-f16.gguf"
],
"strip_prefix": "GigaAM-ASR-GGUF"
},
{
"id": "gigaam_multilingual_ctc_f32",
"display_name": "GigaAM Multilingual CTC F32 GGUF",
"format": "gguf",
"precision": "f32",
"target_directory": "GigaAM-ASR-GGUF",
"files": [
"GigaAM-ASR-GGUF/gigaam-multilingual-ctc-f32.gguf"
],
"strip_prefix": "GigaAM-ASR-GGUF"
},
{
"id": "gigaam_multilingual_ctc_f16",
"display_name": "GigaAM Multilingual CTC F16 GGUF",
"format": "gguf",
"precision": "f16",
"target_directory": "GigaAM-ASR-GGUF",
"files": [
"GigaAM-ASR-GGUF/gigaam-multilingual-ctc-f16.gguf"
],
"strip_prefix": "GigaAM-ASR-GGUF"
},
{
"id": "gigaam_multilingual_large_ctc_f32",
"display_name": "GigaAM Multilingual Large CTC F32 GGUF",
"format": "gguf",
"precision": "f32",
"target_directory": "GigaAM-ASR-GGUF",
"files": [
"GigaAM-ASR-GGUF/gigaam-multilingual-large-ctc-f32.gguf"
],
"strip_prefix": "GigaAM-ASR-GGUF"
},
{
"id": "gigaam_multilingual_large_ctc_f16",
"display_name": "GigaAM Multilingual Large CTC F16 GGUF",
"format": "gguf",
"precision": "f16",
"target_directory": "GigaAM-ASR-GGUF",
"files": [
"GigaAM-ASR-GGUF/gigaam-multilingual-large-ctc-f16.gguf"
],
"strip_prefix": "GigaAM-ASR-GGUF"
}
],
"sources": [
{
"format": "safetensors",
"roots": {
"model": "."
},
"files": {
"config": "model:config.json"
},
"tensors": {
"weights": "model:model.safetensors"
}
},
{
"format": "gguf",
"roots": {
"model": ".",
"weights": "$gguf"
},
"files": {
"config": "model:config.json"
},
"tensors": {
"weights": "weights:"
}
}
]
}