upload Musci-ASR-2.4B
Browse files- .gitattributes +1 -2
- README.md +71 -0
- added_tokens.json +28 -0
- chat_template.jinja +89 -0
- chat_template_default.py +56 -0
- config.json +144 -0
- generation_config.json +4 -0
- merges.txt +0 -0
- model-00000-of-00001.safetensors +3 -0
- model.safetensors.index.json +845 -0
- modeling_Musci.py +333 -0
- processing_Musci.py +219 -0
- special_tokens_map.json +31 -0
- tokenizer.json +3 -0
- tokenizer_config.json +239 -0
- vocab.json +0 -0
.gitattributes
CHANGED
|
@@ -3,12 +3,10 @@
|
|
| 3 |
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 4 |
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
| 5 |
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
| 6 |
-
*.ftz filter=lfs diff=lfs merge=lfs -text
|
| 7 |
*.gz filter=lfs diff=lfs merge=lfs -text
|
| 8 |
*.h5 filter=lfs diff=lfs merge=lfs -text
|
| 9 |
*.joblib filter=lfs diff=lfs merge=lfs -text
|
| 10 |
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
| 11 |
-
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
| 12 |
*.model filter=lfs diff=lfs merge=lfs -text
|
| 13 |
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
| 14 |
*.npy filter=lfs diff=lfs merge=lfs -text
|
|
@@ -33,3 +31,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 3 |
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 4 |
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
| 5 |
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
|
|
|
| 6 |
*.gz filter=lfs diff=lfs merge=lfs -text
|
| 7 |
*.h5 filter=lfs diff=lfs merge=lfs -text
|
| 8 |
*.joblib filter=lfs diff=lfs merge=lfs -text
|
| 9 |
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
| 10 |
*.model filter=lfs diff=lfs merge=lfs -text
|
| 11 |
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
| 12 |
*.npy filter=lfs diff=lfs merge=lfs -text
|
|
|
|
| 31 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 32 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 33 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 34 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language: en
|
| 3 |
+
library_name: transformers
|
| 4 |
+
pipeline_tag: automatic-speech-recognition
|
| 5 |
+
tags:
|
| 6 |
+
- asr
|
| 7 |
+
- speech
|
| 8 |
+
- english
|
| 9 |
+
license: apache-2.0
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Musci-ASR-2.4B
|
| 13 |
+
|
| 14 |
+
An English speech-to-text model that pairs a Qwen3 language-model backbone with a
|
| 15 |
+
Qwen3-Omni-MoE audio encoder. Trained on public English ASR corpora and tuned with
|
| 16 |
+
reinforcement learning on the Open ASR Leaderboard training splits. Total ~2.4B parameters,
|
| 17 |
+
distributed as a single `bfloat16` safetensors shard (~4.84 GB).
|
| 18 |
+
|
| 19 |
+
Submitted to the
|
| 20 |
+
[Open ASR Leaderboard](https://huggingface.co/spaces/hf-audio/open_asr_leaderboard).
|
| 21 |
+
|
| 22 |
+
## Inference
|
| 23 |
+
|
| 24 |
+
```python
|
| 25 |
+
import librosa
|
| 26 |
+
import torch
|
| 27 |
+
from huggingface_hub import hf_hub_download
|
| 28 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
| 29 |
+
from transformers.dynamic_module_utils import get_class_from_dynamic_module
|
| 30 |
+
|
| 31 |
+
REPO = "Musci-research/Musci-ASR-2.4B"
|
| 32 |
+
DEVICE = "cuda:0"
|
| 33 |
+
|
| 34 |
+
model = AutoModelForCausalLM.from_pretrained(
|
| 35 |
+
REPO, torch_dtype=torch.bfloat16, trust_remote_code=True
|
| 36 |
+
).to(DEVICE).eval()
|
| 37 |
+
tokenizer = AutoTokenizer.from_pretrained(REPO, trust_remote_code=True)
|
| 38 |
+
|
| 39 |
+
MusciProcessor = get_class_from_dynamic_module("processing_Musci.MusciProcessor", REPO)
|
| 40 |
+
MelConfig = get_class_from_dynamic_module("processing_Musci.MelConfig", REPO)
|
| 41 |
+
mel_cfg = MelConfig(mel_sr=16000, mel_dim=128, mel_n_fft=400, mel_hop_length=160)
|
| 42 |
+
processor = MusciProcessor(tokenizer, config=mel_cfg, enable_time_marker=False)
|
| 43 |
+
processor.load_template(hf_hub_download(REPO, "chat_template_default.py"))
|
| 44 |
+
|
| 45 |
+
waveform, _ = librosa.load("your_audio.wav", sr=16000)
|
| 46 |
+
inputs = processor(audio=waveform, return_tensors="pt").to(DEVICE)
|
| 47 |
+
inputs["audio_data"] = inputs["audio_data"].to(model.dtype)
|
| 48 |
+
|
| 49 |
+
with torch.no_grad():
|
| 50 |
+
out_ids = model.generate(
|
| 51 |
+
**inputs,
|
| 52 |
+
max_new_tokens=512,
|
| 53 |
+
do_sample=False,
|
| 54 |
+
num_beams=1,
|
| 55 |
+
use_cache=True,
|
| 56 |
+
eos_token_id=[processor.end_token_id],
|
| 57 |
+
)
|
| 58 |
+
|
| 59 |
+
new_ids = out_ids[:, inputs["input_ids"].shape[1]:]
|
| 60 |
+
transcript = processor.batch_decode(new_ids, skip_special_tokens=True)[0].strip()
|
| 61 |
+
print(transcript)
|
| 62 |
+
```
|
| 63 |
+
|
| 64 |
+
## Audio frontend
|
| 65 |
+
|
| 66 |
+
- Sample rate: **16 kHz**
|
| 67 |
+
- Features: Whisper log-mel filterbank — `n_mels=128`, `n_fft=400`, `hop_length=160`
|
| 68 |
+
|
| 69 |
+
## License
|
| 70 |
+
|
| 71 |
+
apache-2.0.
|
added_tokens.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"</think>": 151668,
|
| 3 |
+
"</tool_call>": 151658,
|
| 4 |
+
"</tool_response>": 151666,
|
| 5 |
+
"<think>": 151667,
|
| 6 |
+
"<tool_call>": 151657,
|
| 7 |
+
"<tool_response>": 151665,
|
| 8 |
+
"<|box_end|>": 151649,
|
| 9 |
+
"<|box_start|>": 151648,
|
| 10 |
+
"<|endoftext|>": 151643,
|
| 11 |
+
"<|file_sep|>": 151664,
|
| 12 |
+
"<|fim_middle|>": 151660,
|
| 13 |
+
"<|fim_pad|>": 151662,
|
| 14 |
+
"<|fim_prefix|>": 151659,
|
| 15 |
+
"<|fim_suffix|>": 151661,
|
| 16 |
+
"<|im_end|>": 151645,
|
| 17 |
+
"<|im_start|>": 151644,
|
| 18 |
+
"<|image_pad|>": 151655,
|
| 19 |
+
"<|object_ref_end|>": 151647,
|
| 20 |
+
"<|object_ref_start|>": 151646,
|
| 21 |
+
"<|quad_end|>": 151651,
|
| 22 |
+
"<|quad_start|>": 151650,
|
| 23 |
+
"<|repo_name|>": 151663,
|
| 24 |
+
"<|video_pad|>": 151656,
|
| 25 |
+
"<|vision_end|>": 151653,
|
| 26 |
+
"<|vision_pad|>": 151654,
|
| 27 |
+
"<|vision_start|>": 151652
|
| 28 |
+
}
|
chat_template.jinja
ADDED
|
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0].role == 'system' %}
|
| 4 |
+
{{- messages[0].content + '\n\n' }}
|
| 5 |
+
{%- endif %}
|
| 6 |
+
{{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 7 |
+
{%- for tool in tools %}
|
| 8 |
+
{{- "\n" }}
|
| 9 |
+
{{- tool | tojson }}
|
| 10 |
+
{%- endfor %}
|
| 11 |
+
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
| 12 |
+
{%- else %}
|
| 13 |
+
{%- if messages[0].role == 'system' %}
|
| 14 |
+
{{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
|
| 15 |
+
{%- endif %}
|
| 16 |
+
{%- endif %}
|
| 17 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 18 |
+
{%- for message in messages[::-1] %}
|
| 19 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 20 |
+
{%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
|
| 21 |
+
{%- set ns.multi_step_tool = false %}
|
| 22 |
+
{%- set ns.last_query_index = index %}
|
| 23 |
+
{%- endif %}
|
| 24 |
+
{%- endfor %}
|
| 25 |
+
{%- for message in messages %}
|
| 26 |
+
{%- if message.content is string %}
|
| 27 |
+
{%- set content = message.content %}
|
| 28 |
+
{%- else %}
|
| 29 |
+
{%- set content = '' %}
|
| 30 |
+
{%- endif %}
|
| 31 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
|
| 32 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 33 |
+
{%- elif message.role == "assistant" %}
|
| 34 |
+
{%- set reasoning_content = '' %}
|
| 35 |
+
{%- if message.reasoning_content is string %}
|
| 36 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 37 |
+
{%- else %}
|
| 38 |
+
{%- if '</think>' in content %}
|
| 39 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 40 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 41 |
+
{%- endif %}
|
| 42 |
+
{%- endif %}
|
| 43 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 44 |
+
{%- if loop.last or (not loop.last and reasoning_content) %}
|
| 45 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
|
| 46 |
+
{%- else %}
|
| 47 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 48 |
+
{%- endif %}
|
| 49 |
+
{%- else %}
|
| 50 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 51 |
+
{%- endif %}
|
| 52 |
+
{%- if message.tool_calls %}
|
| 53 |
+
{%- for tool_call in message.tool_calls %}
|
| 54 |
+
{%- if (loop.first and content) or (not loop.first) %}
|
| 55 |
+
{{- '\n' }}
|
| 56 |
+
{%- endif %}
|
| 57 |
+
{%- if tool_call.function %}
|
| 58 |
+
{%- set tool_call = tool_call.function %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<tool_call>\n{"name": "' }}
|
| 61 |
+
{{- tool_call.name }}
|
| 62 |
+
{{- '", "arguments": ' }}
|
| 63 |
+
{%- if tool_call.arguments is string %}
|
| 64 |
+
{{- tool_call.arguments }}
|
| 65 |
+
{%- else %}
|
| 66 |
+
{{- tool_call.arguments | tojson }}
|
| 67 |
+
{%- endif %}
|
| 68 |
+
{{- '}\n</tool_call>' }}
|
| 69 |
+
{%- endfor %}
|
| 70 |
+
{%- endif %}
|
| 71 |
+
{{- '<|im_end|>\n' }}
|
| 72 |
+
{%- elif message.role == "tool" %}
|
| 73 |
+
{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
|
| 74 |
+
{{- '<|im_start|>user' }}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{{- '\n<tool_response>\n' }}
|
| 77 |
+
{{- content }}
|
| 78 |
+
{{- '\n</tool_response>' }}
|
| 79 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 80 |
+
{{- '<|im_end|>\n' }}
|
| 81 |
+
{%- endif %}
|
| 82 |
+
{%- endif %}
|
| 83 |
+
{%- endfor %}
|
| 84 |
+
{%- if add_generation_prompt %}
|
| 85 |
+
{{- '<|im_start|>assistant\n' }}
|
| 86 |
+
{%- if enable_thinking is defined and enable_thinking is false %}
|
| 87 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 88 |
+
{%- endif %}
|
| 89 |
+
{%- endif %}
|
chat_template_default.py
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from dataclasses import dataclass
|
| 2 |
+
from typing import Literal, Optional
|
| 3 |
+
import numpy
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
@dataclass
|
| 7 |
+
class MusciChatTemplateSegment:
|
| 8 |
+
type: Literal["constant_text_token", "text_token", "audio_token", "audio_contiguous"]
|
| 9 |
+
add_loss: bool = True
|
| 10 |
+
text_ids: Optional[numpy.ndarray] = None
|
| 11 |
+
text_token_idx: Optional[int] = None
|
| 12 |
+
text_token_key: Optional[str] = None
|
| 13 |
+
|
| 14 |
+
def __post_init__(self) -> None:
|
| 15 |
+
if self.type == "constant_text_token":
|
| 16 |
+
assert self.text_ids is not None
|
| 17 |
+
elif self.type == "text_token":
|
| 18 |
+
assert self.text_token_key is not None and self.text_token_idx is not None
|
| 19 |
+
elif self.type in ("audio_token", "audio_contiguous"):
|
| 20 |
+
assert not self.add_loss
|
| 21 |
+
|
| 22 |
+
|
| 23 |
+
STYLE_CONTROL_TEXT = ""
|
| 24 |
+
|
| 25 |
+
chat_template = [
|
| 26 |
+
# <|im_start|>user\n<|audio_start|>
|
| 27 |
+
MusciChatTemplateSegment(
|
| 28 |
+
type="constant_text_token",
|
| 29 |
+
text_ids=numpy.array([151644, 872, 198, 151669]),
|
| 30 |
+
add_loss=False,
|
| 31 |
+
),
|
| 32 |
+
MusciChatTemplateSegment(
|
| 33 |
+
type="audio_contiguous",
|
| 34 |
+
add_loss=False,
|
| 35 |
+
),
|
| 36 |
+
# <|audio_end|><|im_end|>\n<|im_start|>assistant\n
|
| 37 |
+
MusciChatTemplateSegment(
|
| 38 |
+
type="constant_text_token",
|
| 39 |
+
text_ids=numpy.array([151670, 151645, 198, 151644, 77091, 198]),
|
| 40 |
+
add_loss=False,
|
| 41 |
+
),
|
| 42 |
+
MusciChatTemplateSegment(
|
| 43 |
+
type="text_token",
|
| 44 |
+
text_token_key="text_token_transcript",
|
| 45 |
+
text_token_idx=0,
|
| 46 |
+
add_loss=True,
|
| 47 |
+
),
|
| 48 |
+
# <|im_end|>
|
| 49 |
+
MusciChatTemplateSegment(
|
| 50 |
+
type="constant_text_token",
|
| 51 |
+
text_ids=numpy.array([151645]),
|
| 52 |
+
add_loss=True,
|
| 53 |
+
),
|
| 54 |
+
]
|
| 55 |
+
|
| 56 |
+
__all__ = ["MusciChatTemplateSegment", "STYLE_CONTROL_TEXT", "chat_template"]
|
config.json
ADDED
|
@@ -0,0 +1,144 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"adapter_hidden_size": 8192,
|
| 3 |
+
"architectures": [
|
| 4 |
+
"MusciForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"audio_config": {
|
| 9 |
+
"activation_dropout": 0,
|
| 10 |
+
"activation_function": "gelu",
|
| 11 |
+
"attention_dropout": 0,
|
| 12 |
+
"conv_chunksize": 500,
|
| 13 |
+
"d_model": 1280,
|
| 14 |
+
"downsample_hidden_size": 480,
|
| 15 |
+
"dropout": 0,
|
| 16 |
+
"encoder_attention_heads": 20,
|
| 17 |
+
"encoder_ffn_dim": 5120,
|
| 18 |
+
"encoder_layers": 32,
|
| 19 |
+
"initializer_range": 0.02,
|
| 20 |
+
"max_source_positions": 1500,
|
| 21 |
+
"model_type": "qwen3_omni_moe_audio_encoder",
|
| 22 |
+
"n_window": 50,
|
| 23 |
+
"n_window_infer": 800,
|
| 24 |
+
"num_hidden_layers": 32,
|
| 25 |
+
"num_mel_bins": 128,
|
| 26 |
+
"output_dim": 2048,
|
| 27 |
+
"scale_embedding": false
|
| 28 |
+
},
|
| 29 |
+
"dtype": "bfloat16",
|
| 30 |
+
"head_dim": 128,
|
| 31 |
+
"hidden_act": "silu",
|
| 32 |
+
"hidden_size": 4096,
|
| 33 |
+
"ignore_index": -100,
|
| 34 |
+
"initializer_range": 0.02,
|
| 35 |
+
"intermediate_size": 22016,
|
| 36 |
+
"language_config": {
|
| 37 |
+
"architectures": [
|
| 38 |
+
"Qwen3ForCausalLM"
|
| 39 |
+
],
|
| 40 |
+
"attention_bias": false,
|
| 41 |
+
"attention_dropout": 0.0,
|
| 42 |
+
"bos_token_id": 151643,
|
| 43 |
+
"dtype": "bfloat16",
|
| 44 |
+
"eos_token_id": 151645,
|
| 45 |
+
"head_dim": 128,
|
| 46 |
+
"hidden_act": "silu",
|
| 47 |
+
"hidden_size": 2048,
|
| 48 |
+
"initializer_range": 0.02,
|
| 49 |
+
"intermediate_size": 6144,
|
| 50 |
+
"layer_types": [
|
| 51 |
+
"full_attention",
|
| 52 |
+
"full_attention",
|
| 53 |
+
"full_attention",
|
| 54 |
+
"full_attention",
|
| 55 |
+
"full_attention",
|
| 56 |
+
"full_attention",
|
| 57 |
+
"full_attention",
|
| 58 |
+
"full_attention",
|
| 59 |
+
"full_attention",
|
| 60 |
+
"full_attention",
|
| 61 |
+
"full_attention",
|
| 62 |
+
"full_attention",
|
| 63 |
+
"full_attention",
|
| 64 |
+
"full_attention",
|
| 65 |
+
"full_attention",
|
| 66 |
+
"full_attention",
|
| 67 |
+
"full_attention",
|
| 68 |
+
"full_attention",
|
| 69 |
+
"full_attention",
|
| 70 |
+
"full_attention",
|
| 71 |
+
"full_attention",
|
| 72 |
+
"full_attention",
|
| 73 |
+
"full_attention",
|
| 74 |
+
"full_attention",
|
| 75 |
+
"full_attention",
|
| 76 |
+
"full_attention",
|
| 77 |
+
"full_attention",
|
| 78 |
+
"full_attention"
|
| 79 |
+
],
|
| 80 |
+
"max_position_embeddings": 40960,
|
| 81 |
+
"max_window_layers": 28,
|
| 82 |
+
"model_type": "qwen3",
|
| 83 |
+
"num_attention_heads": 16,
|
| 84 |
+
"num_hidden_layers": 28,
|
| 85 |
+
"num_key_value_heads": 8,
|
| 86 |
+
"rms_norm_eps": 1e-06,
|
| 87 |
+
"rope_scaling": null,
|
| 88 |
+
"rope_theta": 1000000,
|
| 89 |
+
"sliding_window": null,
|
| 90 |
+
"tie_word_embeddings": true,
|
| 91 |
+
"use_cache": true,
|
| 92 |
+
"use_sliding_window": false,
|
| 93 |
+
"vocab_size": 151936
|
| 94 |
+
},
|
| 95 |
+
"layer_types": [
|
| 96 |
+
"full_attention",
|
| 97 |
+
"full_attention",
|
| 98 |
+
"full_attention",
|
| 99 |
+
"full_attention",
|
| 100 |
+
"full_attention",
|
| 101 |
+
"full_attention",
|
| 102 |
+
"full_attention",
|
| 103 |
+
"full_attention",
|
| 104 |
+
"full_attention",
|
| 105 |
+
"full_attention",
|
| 106 |
+
"full_attention",
|
| 107 |
+
"full_attention",
|
| 108 |
+
"full_attention",
|
| 109 |
+
"full_attention",
|
| 110 |
+
"full_attention",
|
| 111 |
+
"full_attention",
|
| 112 |
+
"full_attention",
|
| 113 |
+
"full_attention",
|
| 114 |
+
"full_attention",
|
| 115 |
+
"full_attention",
|
| 116 |
+
"full_attention",
|
| 117 |
+
"full_attention",
|
| 118 |
+
"full_attention",
|
| 119 |
+
"full_attention",
|
| 120 |
+
"full_attention",
|
| 121 |
+
"full_attention",
|
| 122 |
+
"full_attention",
|
| 123 |
+
"full_attention"
|
| 124 |
+
],
|
| 125 |
+
"max_position_embeddings": 32768,
|
| 126 |
+
"max_window_layers": 28,
|
| 127 |
+
"model_type": "musci",
|
| 128 |
+
"num_attention_heads": 32,
|
| 129 |
+
"num_hidden_layers": 28,
|
| 130 |
+
"num_key_value_heads": 32,
|
| 131 |
+
"rms_norm_eps": 1e-06,
|
| 132 |
+
"rope_scaling": null,
|
| 133 |
+
"rope_theta": 10000.0,
|
| 134 |
+
"sliding_window": null,
|
| 135 |
+
"tie_word_embeddings": false,
|
| 136 |
+
"transformers_version": "4.57.1",
|
| 137 |
+
"use_cache": true,
|
| 138 |
+
"use_sliding_window": false,
|
| 139 |
+
"vocab_size": 151936,
|
| 140 |
+
"auto_map": {
|
| 141 |
+
"AutoConfig": "modeling_Musci.MusciConfig",
|
| 142 |
+
"AutoModelForCausalLM": "modeling_Musci.MusciForCausalLM"
|
| 143 |
+
}
|
| 144 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"transformers_version": "4.57.1"
|
| 4 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
model-00000-of-00001.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7927b459a2c3132883a29252d7f0464fe352cf5abb3c35d2d887a1b5b9d7adae
|
| 3 |
+
size 4837771952
|
model.safetensors.index.json
ADDED
|
@@ -0,0 +1,845 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"metadata": {
|
| 3 |
+
"total_size": 4837667584
|
| 4 |
+
},
|
| 5 |
+
"weight_map": {
|
| 6 |
+
"model.audio_adapter.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 7 |
+
"model.audio_adapter.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 8 |
+
"model.audio_adapter.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 9 |
+
"model.audio_model.conv2d1.bias": "model-00000-of-00001.safetensors",
|
| 10 |
+
"model.audio_model.conv2d1.weight": "model-00000-of-00001.safetensors",
|
| 11 |
+
"model.audio_model.conv2d2.bias": "model-00000-of-00001.safetensors",
|
| 12 |
+
"model.audio_model.conv2d2.weight": "model-00000-of-00001.safetensors",
|
| 13 |
+
"model.audio_model.conv2d3.bias": "model-00000-of-00001.safetensors",
|
| 14 |
+
"model.audio_model.conv2d3.weight": "model-00000-of-00001.safetensors",
|
| 15 |
+
"model.audio_model.conv_out.weight": "model-00000-of-00001.safetensors",
|
| 16 |
+
"model.audio_model.layers.0.fc1.bias": "model-00000-of-00001.safetensors",
|
| 17 |
+
"model.audio_model.layers.0.fc1.weight": "model-00000-of-00001.safetensors",
|
| 18 |
+
"model.audio_model.layers.0.fc2.bias": "model-00000-of-00001.safetensors",
|
| 19 |
+
"model.audio_model.layers.0.fc2.weight": "model-00000-of-00001.safetensors",
|
| 20 |
+
"model.audio_model.layers.0.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 21 |
+
"model.audio_model.layers.0.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 22 |
+
"model.audio_model.layers.0.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 23 |
+
"model.audio_model.layers.0.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 24 |
+
"model.audio_model.layers.0.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 25 |
+
"model.audio_model.layers.0.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 26 |
+
"model.audio_model.layers.0.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 27 |
+
"model.audio_model.layers.0.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 28 |
+
"model.audio_model.layers.0.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 29 |
+
"model.audio_model.layers.0.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 30 |
+
"model.audio_model.layers.0.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 31 |
+
"model.audio_model.layers.0.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 32 |
+
"model.audio_model.layers.1.fc1.bias": "model-00000-of-00001.safetensors",
|
| 33 |
+
"model.audio_model.layers.1.fc1.weight": "model-00000-of-00001.safetensors",
|
| 34 |
+
"model.audio_model.layers.1.fc2.bias": "model-00000-of-00001.safetensors",
|
| 35 |
+
"model.audio_model.layers.1.fc2.weight": "model-00000-of-00001.safetensors",
|
| 36 |
+
"model.audio_model.layers.1.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 37 |
+
"model.audio_model.layers.1.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 38 |
+
"model.audio_model.layers.1.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 39 |
+
"model.audio_model.layers.1.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 40 |
+
"model.audio_model.layers.1.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 41 |
+
"model.audio_model.layers.1.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 42 |
+
"model.audio_model.layers.1.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 43 |
+
"model.audio_model.layers.1.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 44 |
+
"model.audio_model.layers.1.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 45 |
+
"model.audio_model.layers.1.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 46 |
+
"model.audio_model.layers.1.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 47 |
+
"model.audio_model.layers.1.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 48 |
+
"model.audio_model.layers.10.fc1.bias": "model-00000-of-00001.safetensors",
|
| 49 |
+
"model.audio_model.layers.10.fc1.weight": "model-00000-of-00001.safetensors",
|
| 50 |
+
"model.audio_model.layers.10.fc2.bias": "model-00000-of-00001.safetensors",
|
| 51 |
+
"model.audio_model.layers.10.fc2.weight": "model-00000-of-00001.safetensors",
|
| 52 |
+
"model.audio_model.layers.10.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 53 |
+
"model.audio_model.layers.10.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 54 |
+
"model.audio_model.layers.10.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 55 |
+
"model.audio_model.layers.10.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 56 |
+
"model.audio_model.layers.10.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 57 |
+
"model.audio_model.layers.10.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 58 |
+
"model.audio_model.layers.10.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 59 |
+
"model.audio_model.layers.10.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 60 |
+
"model.audio_model.layers.10.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 61 |
+
"model.audio_model.layers.10.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 62 |
+
"model.audio_model.layers.10.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 63 |
+
"model.audio_model.layers.10.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 64 |
+
"model.audio_model.layers.11.fc1.bias": "model-00000-of-00001.safetensors",
|
| 65 |
+
"model.audio_model.layers.11.fc1.weight": "model-00000-of-00001.safetensors",
|
| 66 |
+
"model.audio_model.layers.11.fc2.bias": "model-00000-of-00001.safetensors",
|
| 67 |
+
"model.audio_model.layers.11.fc2.weight": "model-00000-of-00001.safetensors",
|
| 68 |
+
"model.audio_model.layers.11.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 69 |
+
"model.audio_model.layers.11.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 70 |
+
"model.audio_model.layers.11.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 71 |
+
"model.audio_model.layers.11.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 72 |
+
"model.audio_model.layers.11.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 73 |
+
"model.audio_model.layers.11.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 74 |
+
"model.audio_model.layers.11.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 75 |
+
"model.audio_model.layers.11.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 76 |
+
"model.audio_model.layers.11.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 77 |
+
"model.audio_model.layers.11.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 78 |
+
"model.audio_model.layers.11.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 79 |
+
"model.audio_model.layers.11.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 80 |
+
"model.audio_model.layers.12.fc1.bias": "model-00000-of-00001.safetensors",
|
| 81 |
+
"model.audio_model.layers.12.fc1.weight": "model-00000-of-00001.safetensors",
|
| 82 |
+
"model.audio_model.layers.12.fc2.bias": "model-00000-of-00001.safetensors",
|
| 83 |
+
"model.audio_model.layers.12.fc2.weight": "model-00000-of-00001.safetensors",
|
| 84 |
+
"model.audio_model.layers.12.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 85 |
+
"model.audio_model.layers.12.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 86 |
+
"model.audio_model.layers.12.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 87 |
+
"model.audio_model.layers.12.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 88 |
+
"model.audio_model.layers.12.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 89 |
+
"model.audio_model.layers.12.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 90 |
+
"model.audio_model.layers.12.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 91 |
+
"model.audio_model.layers.12.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 92 |
+
"model.audio_model.layers.12.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 93 |
+
"model.audio_model.layers.12.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 94 |
+
"model.audio_model.layers.12.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 95 |
+
"model.audio_model.layers.12.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 96 |
+
"model.audio_model.layers.13.fc1.bias": "model-00000-of-00001.safetensors",
|
| 97 |
+
"model.audio_model.layers.13.fc1.weight": "model-00000-of-00001.safetensors",
|
| 98 |
+
"model.audio_model.layers.13.fc2.bias": "model-00000-of-00001.safetensors",
|
| 99 |
+
"model.audio_model.layers.13.fc2.weight": "model-00000-of-00001.safetensors",
|
| 100 |
+
"model.audio_model.layers.13.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 101 |
+
"model.audio_model.layers.13.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 102 |
+
"model.audio_model.layers.13.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 103 |
+
"model.audio_model.layers.13.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 104 |
+
"model.audio_model.layers.13.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 105 |
+
"model.audio_model.layers.13.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 106 |
+
"model.audio_model.layers.13.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 107 |
+
"model.audio_model.layers.13.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 108 |
+
"model.audio_model.layers.13.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 109 |
+
"model.audio_model.layers.13.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 110 |
+
"model.audio_model.layers.13.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 111 |
+
"model.audio_model.layers.13.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 112 |
+
"model.audio_model.layers.14.fc1.bias": "model-00000-of-00001.safetensors",
|
| 113 |
+
"model.audio_model.layers.14.fc1.weight": "model-00000-of-00001.safetensors",
|
| 114 |
+
"model.audio_model.layers.14.fc2.bias": "model-00000-of-00001.safetensors",
|
| 115 |
+
"model.audio_model.layers.14.fc2.weight": "model-00000-of-00001.safetensors",
|
| 116 |
+
"model.audio_model.layers.14.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 117 |
+
"model.audio_model.layers.14.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 118 |
+
"model.audio_model.layers.14.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 119 |
+
"model.audio_model.layers.14.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 120 |
+
"model.audio_model.layers.14.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 121 |
+
"model.audio_model.layers.14.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 122 |
+
"model.audio_model.layers.14.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 123 |
+
"model.audio_model.layers.14.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 124 |
+
"model.audio_model.layers.14.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 125 |
+
"model.audio_model.layers.14.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 126 |
+
"model.audio_model.layers.14.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 127 |
+
"model.audio_model.layers.14.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 128 |
+
"model.audio_model.layers.15.fc1.bias": "model-00000-of-00001.safetensors",
|
| 129 |
+
"model.audio_model.layers.15.fc1.weight": "model-00000-of-00001.safetensors",
|
| 130 |
+
"model.audio_model.layers.15.fc2.bias": "model-00000-of-00001.safetensors",
|
| 131 |
+
"model.audio_model.layers.15.fc2.weight": "model-00000-of-00001.safetensors",
|
| 132 |
+
"model.audio_model.layers.15.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 133 |
+
"model.audio_model.layers.15.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 134 |
+
"model.audio_model.layers.15.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 135 |
+
"model.audio_model.layers.15.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 136 |
+
"model.audio_model.layers.15.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 137 |
+
"model.audio_model.layers.15.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 138 |
+
"model.audio_model.layers.15.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 139 |
+
"model.audio_model.layers.15.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 140 |
+
"model.audio_model.layers.15.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 141 |
+
"model.audio_model.layers.15.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 142 |
+
"model.audio_model.layers.15.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 143 |
+
"model.audio_model.layers.15.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 144 |
+
"model.audio_model.layers.16.fc1.bias": "model-00000-of-00001.safetensors",
|
| 145 |
+
"model.audio_model.layers.16.fc1.weight": "model-00000-of-00001.safetensors",
|
| 146 |
+
"model.audio_model.layers.16.fc2.bias": "model-00000-of-00001.safetensors",
|
| 147 |
+
"model.audio_model.layers.16.fc2.weight": "model-00000-of-00001.safetensors",
|
| 148 |
+
"model.audio_model.layers.16.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 149 |
+
"model.audio_model.layers.16.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 150 |
+
"model.audio_model.layers.16.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 151 |
+
"model.audio_model.layers.16.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 152 |
+
"model.audio_model.layers.16.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 153 |
+
"model.audio_model.layers.16.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 154 |
+
"model.audio_model.layers.16.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 155 |
+
"model.audio_model.layers.16.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 156 |
+
"model.audio_model.layers.16.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 157 |
+
"model.audio_model.layers.16.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 158 |
+
"model.audio_model.layers.16.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 159 |
+
"model.audio_model.layers.16.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 160 |
+
"model.audio_model.layers.17.fc1.bias": "model-00000-of-00001.safetensors",
|
| 161 |
+
"model.audio_model.layers.17.fc1.weight": "model-00000-of-00001.safetensors",
|
| 162 |
+
"model.audio_model.layers.17.fc2.bias": "model-00000-of-00001.safetensors",
|
| 163 |
+
"model.audio_model.layers.17.fc2.weight": "model-00000-of-00001.safetensors",
|
| 164 |
+
"model.audio_model.layers.17.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 165 |
+
"model.audio_model.layers.17.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 166 |
+
"model.audio_model.layers.17.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 167 |
+
"model.audio_model.layers.17.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 168 |
+
"model.audio_model.layers.17.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 169 |
+
"model.audio_model.layers.17.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 170 |
+
"model.audio_model.layers.17.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 171 |
+
"model.audio_model.layers.17.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 172 |
+
"model.audio_model.layers.17.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 173 |
+
"model.audio_model.layers.17.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 174 |
+
"model.audio_model.layers.17.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 175 |
+
"model.audio_model.layers.17.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 176 |
+
"model.audio_model.layers.18.fc1.bias": "model-00000-of-00001.safetensors",
|
| 177 |
+
"model.audio_model.layers.18.fc1.weight": "model-00000-of-00001.safetensors",
|
| 178 |
+
"model.audio_model.layers.18.fc2.bias": "model-00000-of-00001.safetensors",
|
| 179 |
+
"model.audio_model.layers.18.fc2.weight": "model-00000-of-00001.safetensors",
|
| 180 |
+
"model.audio_model.layers.18.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 181 |
+
"model.audio_model.layers.18.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 182 |
+
"model.audio_model.layers.18.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 183 |
+
"model.audio_model.layers.18.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 184 |
+
"model.audio_model.layers.18.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 185 |
+
"model.audio_model.layers.18.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 186 |
+
"model.audio_model.layers.18.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 187 |
+
"model.audio_model.layers.18.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 188 |
+
"model.audio_model.layers.18.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 189 |
+
"model.audio_model.layers.18.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 190 |
+
"model.audio_model.layers.18.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 191 |
+
"model.audio_model.layers.18.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 192 |
+
"model.audio_model.layers.19.fc1.bias": "model-00000-of-00001.safetensors",
|
| 193 |
+
"model.audio_model.layers.19.fc1.weight": "model-00000-of-00001.safetensors",
|
| 194 |
+
"model.audio_model.layers.19.fc2.bias": "model-00000-of-00001.safetensors",
|
| 195 |
+
"model.audio_model.layers.19.fc2.weight": "model-00000-of-00001.safetensors",
|
| 196 |
+
"model.audio_model.layers.19.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 197 |
+
"model.audio_model.layers.19.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 198 |
+
"model.audio_model.layers.19.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 199 |
+
"model.audio_model.layers.19.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 200 |
+
"model.audio_model.layers.19.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 201 |
+
"model.audio_model.layers.19.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 202 |
+
"model.audio_model.layers.19.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 203 |
+
"model.audio_model.layers.19.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 204 |
+
"model.audio_model.layers.19.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 205 |
+
"model.audio_model.layers.19.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 206 |
+
"model.audio_model.layers.19.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 207 |
+
"model.audio_model.layers.19.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 208 |
+
"model.audio_model.layers.2.fc1.bias": "model-00000-of-00001.safetensors",
|
| 209 |
+
"model.audio_model.layers.2.fc1.weight": "model-00000-of-00001.safetensors",
|
| 210 |
+
"model.audio_model.layers.2.fc2.bias": "model-00000-of-00001.safetensors",
|
| 211 |
+
"model.audio_model.layers.2.fc2.weight": "model-00000-of-00001.safetensors",
|
| 212 |
+
"model.audio_model.layers.2.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 213 |
+
"model.audio_model.layers.2.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 214 |
+
"model.audio_model.layers.2.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 215 |
+
"model.audio_model.layers.2.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 216 |
+
"model.audio_model.layers.2.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 217 |
+
"model.audio_model.layers.2.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 218 |
+
"model.audio_model.layers.2.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 219 |
+
"model.audio_model.layers.2.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 220 |
+
"model.audio_model.layers.2.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 221 |
+
"model.audio_model.layers.2.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 222 |
+
"model.audio_model.layers.2.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 223 |
+
"model.audio_model.layers.2.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 224 |
+
"model.audio_model.layers.20.fc1.bias": "model-00000-of-00001.safetensors",
|
| 225 |
+
"model.audio_model.layers.20.fc1.weight": "model-00000-of-00001.safetensors",
|
| 226 |
+
"model.audio_model.layers.20.fc2.bias": "model-00000-of-00001.safetensors",
|
| 227 |
+
"model.audio_model.layers.20.fc2.weight": "model-00000-of-00001.safetensors",
|
| 228 |
+
"model.audio_model.layers.20.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 229 |
+
"model.audio_model.layers.20.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 230 |
+
"model.audio_model.layers.20.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 231 |
+
"model.audio_model.layers.20.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 232 |
+
"model.audio_model.layers.20.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 233 |
+
"model.audio_model.layers.20.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 234 |
+
"model.audio_model.layers.20.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 235 |
+
"model.audio_model.layers.20.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 236 |
+
"model.audio_model.layers.20.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 237 |
+
"model.audio_model.layers.20.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 238 |
+
"model.audio_model.layers.20.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 239 |
+
"model.audio_model.layers.20.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 240 |
+
"model.audio_model.layers.21.fc1.bias": "model-00000-of-00001.safetensors",
|
| 241 |
+
"model.audio_model.layers.21.fc1.weight": "model-00000-of-00001.safetensors",
|
| 242 |
+
"model.audio_model.layers.21.fc2.bias": "model-00000-of-00001.safetensors",
|
| 243 |
+
"model.audio_model.layers.21.fc2.weight": "model-00000-of-00001.safetensors",
|
| 244 |
+
"model.audio_model.layers.21.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 245 |
+
"model.audio_model.layers.21.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 246 |
+
"model.audio_model.layers.21.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 247 |
+
"model.audio_model.layers.21.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 248 |
+
"model.audio_model.layers.21.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 249 |
+
"model.audio_model.layers.21.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 250 |
+
"model.audio_model.layers.21.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 251 |
+
"model.audio_model.layers.21.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 252 |
+
"model.audio_model.layers.21.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 253 |
+
"model.audio_model.layers.21.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 254 |
+
"model.audio_model.layers.21.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 255 |
+
"model.audio_model.layers.21.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 256 |
+
"model.audio_model.layers.22.fc1.bias": "model-00000-of-00001.safetensors",
|
| 257 |
+
"model.audio_model.layers.22.fc1.weight": "model-00000-of-00001.safetensors",
|
| 258 |
+
"model.audio_model.layers.22.fc2.bias": "model-00000-of-00001.safetensors",
|
| 259 |
+
"model.audio_model.layers.22.fc2.weight": "model-00000-of-00001.safetensors",
|
| 260 |
+
"model.audio_model.layers.22.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 261 |
+
"model.audio_model.layers.22.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 262 |
+
"model.audio_model.layers.22.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 263 |
+
"model.audio_model.layers.22.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 264 |
+
"model.audio_model.layers.22.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 265 |
+
"model.audio_model.layers.22.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 266 |
+
"model.audio_model.layers.22.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 267 |
+
"model.audio_model.layers.22.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 268 |
+
"model.audio_model.layers.22.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 269 |
+
"model.audio_model.layers.22.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 270 |
+
"model.audio_model.layers.22.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 271 |
+
"model.audio_model.layers.22.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 272 |
+
"model.audio_model.layers.23.fc1.bias": "model-00000-of-00001.safetensors",
|
| 273 |
+
"model.audio_model.layers.23.fc1.weight": "model-00000-of-00001.safetensors",
|
| 274 |
+
"model.audio_model.layers.23.fc2.bias": "model-00000-of-00001.safetensors",
|
| 275 |
+
"model.audio_model.layers.23.fc2.weight": "model-00000-of-00001.safetensors",
|
| 276 |
+
"model.audio_model.layers.23.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 277 |
+
"model.audio_model.layers.23.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 278 |
+
"model.audio_model.layers.23.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 279 |
+
"model.audio_model.layers.23.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 280 |
+
"model.audio_model.layers.23.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 281 |
+
"model.audio_model.layers.23.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 282 |
+
"model.audio_model.layers.23.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 283 |
+
"model.audio_model.layers.23.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 284 |
+
"model.audio_model.layers.23.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 285 |
+
"model.audio_model.layers.23.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 286 |
+
"model.audio_model.layers.23.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 287 |
+
"model.audio_model.layers.23.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 288 |
+
"model.audio_model.layers.24.fc1.bias": "model-00000-of-00001.safetensors",
|
| 289 |
+
"model.audio_model.layers.24.fc1.weight": "model-00000-of-00001.safetensors",
|
| 290 |
+
"model.audio_model.layers.24.fc2.bias": "model-00000-of-00001.safetensors",
|
| 291 |
+
"model.audio_model.layers.24.fc2.weight": "model-00000-of-00001.safetensors",
|
| 292 |
+
"model.audio_model.layers.24.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 293 |
+
"model.audio_model.layers.24.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 294 |
+
"model.audio_model.layers.24.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 295 |
+
"model.audio_model.layers.24.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 296 |
+
"model.audio_model.layers.24.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 297 |
+
"model.audio_model.layers.24.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 298 |
+
"model.audio_model.layers.24.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 299 |
+
"model.audio_model.layers.24.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 300 |
+
"model.audio_model.layers.24.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 301 |
+
"model.audio_model.layers.24.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 302 |
+
"model.audio_model.layers.24.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 303 |
+
"model.audio_model.layers.24.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 304 |
+
"model.audio_model.layers.25.fc1.bias": "model-00000-of-00001.safetensors",
|
| 305 |
+
"model.audio_model.layers.25.fc1.weight": "model-00000-of-00001.safetensors",
|
| 306 |
+
"model.audio_model.layers.25.fc2.bias": "model-00000-of-00001.safetensors",
|
| 307 |
+
"model.audio_model.layers.25.fc2.weight": "model-00000-of-00001.safetensors",
|
| 308 |
+
"model.audio_model.layers.25.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 309 |
+
"model.audio_model.layers.25.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 310 |
+
"model.audio_model.layers.25.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 311 |
+
"model.audio_model.layers.25.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 312 |
+
"model.audio_model.layers.25.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 313 |
+
"model.audio_model.layers.25.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 314 |
+
"model.audio_model.layers.25.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 315 |
+
"model.audio_model.layers.25.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 316 |
+
"model.audio_model.layers.25.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 317 |
+
"model.audio_model.layers.25.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 318 |
+
"model.audio_model.layers.25.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 319 |
+
"model.audio_model.layers.25.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 320 |
+
"model.audio_model.layers.26.fc1.bias": "model-00000-of-00001.safetensors",
|
| 321 |
+
"model.audio_model.layers.26.fc1.weight": "model-00000-of-00001.safetensors",
|
| 322 |
+
"model.audio_model.layers.26.fc2.bias": "model-00000-of-00001.safetensors",
|
| 323 |
+
"model.audio_model.layers.26.fc2.weight": "model-00000-of-00001.safetensors",
|
| 324 |
+
"model.audio_model.layers.26.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 325 |
+
"model.audio_model.layers.26.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 326 |
+
"model.audio_model.layers.26.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 327 |
+
"model.audio_model.layers.26.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 328 |
+
"model.audio_model.layers.26.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 329 |
+
"model.audio_model.layers.26.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 330 |
+
"model.audio_model.layers.26.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 331 |
+
"model.audio_model.layers.26.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 332 |
+
"model.audio_model.layers.26.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 333 |
+
"model.audio_model.layers.26.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 334 |
+
"model.audio_model.layers.26.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 335 |
+
"model.audio_model.layers.26.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 336 |
+
"model.audio_model.layers.27.fc1.bias": "model-00000-of-00001.safetensors",
|
| 337 |
+
"model.audio_model.layers.27.fc1.weight": "model-00000-of-00001.safetensors",
|
| 338 |
+
"model.audio_model.layers.27.fc2.bias": "model-00000-of-00001.safetensors",
|
| 339 |
+
"model.audio_model.layers.27.fc2.weight": "model-00000-of-00001.safetensors",
|
| 340 |
+
"model.audio_model.layers.27.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 341 |
+
"model.audio_model.layers.27.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 342 |
+
"model.audio_model.layers.27.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 343 |
+
"model.audio_model.layers.27.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 344 |
+
"model.audio_model.layers.27.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 345 |
+
"model.audio_model.layers.27.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 346 |
+
"model.audio_model.layers.27.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 347 |
+
"model.audio_model.layers.27.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 348 |
+
"model.audio_model.layers.27.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 349 |
+
"model.audio_model.layers.27.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 350 |
+
"model.audio_model.layers.27.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 351 |
+
"model.audio_model.layers.27.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 352 |
+
"model.audio_model.layers.28.fc1.bias": "model-00000-of-00001.safetensors",
|
| 353 |
+
"model.audio_model.layers.28.fc1.weight": "model-00000-of-00001.safetensors",
|
| 354 |
+
"model.audio_model.layers.28.fc2.bias": "model-00000-of-00001.safetensors",
|
| 355 |
+
"model.audio_model.layers.28.fc2.weight": "model-00000-of-00001.safetensors",
|
| 356 |
+
"model.audio_model.layers.28.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 357 |
+
"model.audio_model.layers.28.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 358 |
+
"model.audio_model.layers.28.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 359 |
+
"model.audio_model.layers.28.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 360 |
+
"model.audio_model.layers.28.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 361 |
+
"model.audio_model.layers.28.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 362 |
+
"model.audio_model.layers.28.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 363 |
+
"model.audio_model.layers.28.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 364 |
+
"model.audio_model.layers.28.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 365 |
+
"model.audio_model.layers.28.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 366 |
+
"model.audio_model.layers.28.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 367 |
+
"model.audio_model.layers.28.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 368 |
+
"model.audio_model.layers.29.fc1.bias": "model-00000-of-00001.safetensors",
|
| 369 |
+
"model.audio_model.layers.29.fc1.weight": "model-00000-of-00001.safetensors",
|
| 370 |
+
"model.audio_model.layers.29.fc2.bias": "model-00000-of-00001.safetensors",
|
| 371 |
+
"model.audio_model.layers.29.fc2.weight": "model-00000-of-00001.safetensors",
|
| 372 |
+
"model.audio_model.layers.29.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 373 |
+
"model.audio_model.layers.29.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 374 |
+
"model.audio_model.layers.29.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 375 |
+
"model.audio_model.layers.29.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 376 |
+
"model.audio_model.layers.29.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 377 |
+
"model.audio_model.layers.29.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 378 |
+
"model.audio_model.layers.29.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 379 |
+
"model.audio_model.layers.29.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 380 |
+
"model.audio_model.layers.29.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 381 |
+
"model.audio_model.layers.29.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 382 |
+
"model.audio_model.layers.29.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 383 |
+
"model.audio_model.layers.29.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 384 |
+
"model.audio_model.layers.3.fc1.bias": "model-00000-of-00001.safetensors",
|
| 385 |
+
"model.audio_model.layers.3.fc1.weight": "model-00000-of-00001.safetensors",
|
| 386 |
+
"model.audio_model.layers.3.fc2.bias": "model-00000-of-00001.safetensors",
|
| 387 |
+
"model.audio_model.layers.3.fc2.weight": "model-00000-of-00001.safetensors",
|
| 388 |
+
"model.audio_model.layers.3.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 389 |
+
"model.audio_model.layers.3.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 390 |
+
"model.audio_model.layers.3.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 391 |
+
"model.audio_model.layers.3.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 392 |
+
"model.audio_model.layers.3.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 393 |
+
"model.audio_model.layers.3.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 394 |
+
"model.audio_model.layers.3.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 395 |
+
"model.audio_model.layers.3.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 396 |
+
"model.audio_model.layers.3.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 397 |
+
"model.audio_model.layers.3.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 398 |
+
"model.audio_model.layers.3.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 399 |
+
"model.audio_model.layers.3.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 400 |
+
"model.audio_model.layers.30.fc1.bias": "model-00000-of-00001.safetensors",
|
| 401 |
+
"model.audio_model.layers.30.fc1.weight": "model-00000-of-00001.safetensors",
|
| 402 |
+
"model.audio_model.layers.30.fc2.bias": "model-00000-of-00001.safetensors",
|
| 403 |
+
"model.audio_model.layers.30.fc2.weight": "model-00000-of-00001.safetensors",
|
| 404 |
+
"model.audio_model.layers.30.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 405 |
+
"model.audio_model.layers.30.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 406 |
+
"model.audio_model.layers.30.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 407 |
+
"model.audio_model.layers.30.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 408 |
+
"model.audio_model.layers.30.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 409 |
+
"model.audio_model.layers.30.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 410 |
+
"model.audio_model.layers.30.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 411 |
+
"model.audio_model.layers.30.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 412 |
+
"model.audio_model.layers.30.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 413 |
+
"model.audio_model.layers.30.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 414 |
+
"model.audio_model.layers.30.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 415 |
+
"model.audio_model.layers.30.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 416 |
+
"model.audio_model.layers.31.fc1.bias": "model-00000-of-00001.safetensors",
|
| 417 |
+
"model.audio_model.layers.31.fc1.weight": "model-00000-of-00001.safetensors",
|
| 418 |
+
"model.audio_model.layers.31.fc2.bias": "model-00000-of-00001.safetensors",
|
| 419 |
+
"model.audio_model.layers.31.fc2.weight": "model-00000-of-00001.safetensors",
|
| 420 |
+
"model.audio_model.layers.31.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 421 |
+
"model.audio_model.layers.31.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 422 |
+
"model.audio_model.layers.31.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 423 |
+
"model.audio_model.layers.31.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 424 |
+
"model.audio_model.layers.31.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 425 |
+
"model.audio_model.layers.31.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 426 |
+
"model.audio_model.layers.31.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 427 |
+
"model.audio_model.layers.31.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 428 |
+
"model.audio_model.layers.31.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 429 |
+
"model.audio_model.layers.31.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 430 |
+
"model.audio_model.layers.31.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 431 |
+
"model.audio_model.layers.31.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 432 |
+
"model.audio_model.layers.4.fc1.bias": "model-00000-of-00001.safetensors",
|
| 433 |
+
"model.audio_model.layers.4.fc1.weight": "model-00000-of-00001.safetensors",
|
| 434 |
+
"model.audio_model.layers.4.fc2.bias": "model-00000-of-00001.safetensors",
|
| 435 |
+
"model.audio_model.layers.4.fc2.weight": "model-00000-of-00001.safetensors",
|
| 436 |
+
"model.audio_model.layers.4.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 437 |
+
"model.audio_model.layers.4.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 438 |
+
"model.audio_model.layers.4.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 439 |
+
"model.audio_model.layers.4.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 440 |
+
"model.audio_model.layers.4.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 441 |
+
"model.audio_model.layers.4.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 442 |
+
"model.audio_model.layers.4.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 443 |
+
"model.audio_model.layers.4.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 444 |
+
"model.audio_model.layers.4.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 445 |
+
"model.audio_model.layers.4.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 446 |
+
"model.audio_model.layers.4.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 447 |
+
"model.audio_model.layers.4.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 448 |
+
"model.audio_model.layers.5.fc1.bias": "model-00000-of-00001.safetensors",
|
| 449 |
+
"model.audio_model.layers.5.fc1.weight": "model-00000-of-00001.safetensors",
|
| 450 |
+
"model.audio_model.layers.5.fc2.bias": "model-00000-of-00001.safetensors",
|
| 451 |
+
"model.audio_model.layers.5.fc2.weight": "model-00000-of-00001.safetensors",
|
| 452 |
+
"model.audio_model.layers.5.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 453 |
+
"model.audio_model.layers.5.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 454 |
+
"model.audio_model.layers.5.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 455 |
+
"model.audio_model.layers.5.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 456 |
+
"model.audio_model.layers.5.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 457 |
+
"model.audio_model.layers.5.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 458 |
+
"model.audio_model.layers.5.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 459 |
+
"model.audio_model.layers.5.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 460 |
+
"model.audio_model.layers.5.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 461 |
+
"model.audio_model.layers.5.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 462 |
+
"model.audio_model.layers.5.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 463 |
+
"model.audio_model.layers.5.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 464 |
+
"model.audio_model.layers.6.fc1.bias": "model-00000-of-00001.safetensors",
|
| 465 |
+
"model.audio_model.layers.6.fc1.weight": "model-00000-of-00001.safetensors",
|
| 466 |
+
"model.audio_model.layers.6.fc2.bias": "model-00000-of-00001.safetensors",
|
| 467 |
+
"model.audio_model.layers.6.fc2.weight": "model-00000-of-00001.safetensors",
|
| 468 |
+
"model.audio_model.layers.6.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 469 |
+
"model.audio_model.layers.6.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 470 |
+
"model.audio_model.layers.6.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 471 |
+
"model.audio_model.layers.6.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 472 |
+
"model.audio_model.layers.6.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 473 |
+
"model.audio_model.layers.6.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 474 |
+
"model.audio_model.layers.6.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 475 |
+
"model.audio_model.layers.6.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 476 |
+
"model.audio_model.layers.6.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 477 |
+
"model.audio_model.layers.6.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 478 |
+
"model.audio_model.layers.6.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 479 |
+
"model.audio_model.layers.6.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 480 |
+
"model.audio_model.layers.7.fc1.bias": "model-00000-of-00001.safetensors",
|
| 481 |
+
"model.audio_model.layers.7.fc1.weight": "model-00000-of-00001.safetensors",
|
| 482 |
+
"model.audio_model.layers.7.fc2.bias": "model-00000-of-00001.safetensors",
|
| 483 |
+
"model.audio_model.layers.7.fc2.weight": "model-00000-of-00001.safetensors",
|
| 484 |
+
"model.audio_model.layers.7.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 485 |
+
"model.audio_model.layers.7.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 486 |
+
"model.audio_model.layers.7.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 487 |
+
"model.audio_model.layers.7.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 488 |
+
"model.audio_model.layers.7.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 489 |
+
"model.audio_model.layers.7.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 490 |
+
"model.audio_model.layers.7.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 491 |
+
"model.audio_model.layers.7.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 492 |
+
"model.audio_model.layers.7.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 493 |
+
"model.audio_model.layers.7.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 494 |
+
"model.audio_model.layers.7.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 495 |
+
"model.audio_model.layers.7.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 496 |
+
"model.audio_model.layers.8.fc1.bias": "model-00000-of-00001.safetensors",
|
| 497 |
+
"model.audio_model.layers.8.fc1.weight": "model-00000-of-00001.safetensors",
|
| 498 |
+
"model.audio_model.layers.8.fc2.bias": "model-00000-of-00001.safetensors",
|
| 499 |
+
"model.audio_model.layers.8.fc2.weight": "model-00000-of-00001.safetensors",
|
| 500 |
+
"model.audio_model.layers.8.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 501 |
+
"model.audio_model.layers.8.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 502 |
+
"model.audio_model.layers.8.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 503 |
+
"model.audio_model.layers.8.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 504 |
+
"model.audio_model.layers.8.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 505 |
+
"model.audio_model.layers.8.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 506 |
+
"model.audio_model.layers.8.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 507 |
+
"model.audio_model.layers.8.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 508 |
+
"model.audio_model.layers.8.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 509 |
+
"model.audio_model.layers.8.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 510 |
+
"model.audio_model.layers.8.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 511 |
+
"model.audio_model.layers.8.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 512 |
+
"model.audio_model.layers.9.fc1.bias": "model-00000-of-00001.safetensors",
|
| 513 |
+
"model.audio_model.layers.9.fc1.weight": "model-00000-of-00001.safetensors",
|
| 514 |
+
"model.audio_model.layers.9.fc2.bias": "model-00000-of-00001.safetensors",
|
| 515 |
+
"model.audio_model.layers.9.fc2.weight": "model-00000-of-00001.safetensors",
|
| 516 |
+
"model.audio_model.layers.9.final_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 517 |
+
"model.audio_model.layers.9.final_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 518 |
+
"model.audio_model.layers.9.self_attn.k_proj.bias": "model-00000-of-00001.safetensors",
|
| 519 |
+
"model.audio_model.layers.9.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 520 |
+
"model.audio_model.layers.9.self_attn.out_proj.bias": "model-00000-of-00001.safetensors",
|
| 521 |
+
"model.audio_model.layers.9.self_attn.out_proj.weight": "model-00000-of-00001.safetensors",
|
| 522 |
+
"model.audio_model.layers.9.self_attn.q_proj.bias": "model-00000-of-00001.safetensors",
|
| 523 |
+
"model.audio_model.layers.9.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 524 |
+
"model.audio_model.layers.9.self_attn.v_proj.bias": "model-00000-of-00001.safetensors",
|
| 525 |
+
"model.audio_model.layers.9.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 526 |
+
"model.audio_model.layers.9.self_attn_layer_norm.bias": "model-00000-of-00001.safetensors",
|
| 527 |
+
"model.audio_model.layers.9.self_attn_layer_norm.weight": "model-00000-of-00001.safetensors",
|
| 528 |
+
"model.audio_model.ln_post.bias": "model-00000-of-00001.safetensors",
|
| 529 |
+
"model.audio_model.ln_post.weight": "model-00000-of-00001.safetensors",
|
| 530 |
+
"model.audio_model.proj1.bias": "model-00000-of-00001.safetensors",
|
| 531 |
+
"model.audio_model.proj1.weight": "model-00000-of-00001.safetensors",
|
| 532 |
+
"model.audio_model.proj2.bias": "model-00000-of-00001.safetensors",
|
| 533 |
+
"model.audio_model.proj2.weight": "model-00000-of-00001.safetensors",
|
| 534 |
+
"model.language_model.norm.weight": "model-00000-of-00001.safetensors",
|
| 535 |
+
"model.language_model.layers.0.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 536 |
+
"model.language_model.layers.0.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 537 |
+
"model.language_model.layers.0.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 538 |
+
"model.language_model.layers.0.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 539 |
+
"model.language_model.layers.0.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 540 |
+
"model.language_model.layers.0.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 541 |
+
"model.language_model.layers.0.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 542 |
+
"model.language_model.layers.0.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 543 |
+
"model.language_model.layers.0.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 544 |
+
"model.language_model.layers.0.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 545 |
+
"model.language_model.layers.0.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 546 |
+
"model.language_model.layers.1.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 547 |
+
"model.language_model.layers.1.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 548 |
+
"model.language_model.layers.1.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 549 |
+
"model.language_model.layers.1.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 550 |
+
"model.language_model.layers.1.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 551 |
+
"model.language_model.layers.1.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 552 |
+
"model.language_model.layers.1.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 553 |
+
"model.language_model.layers.1.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 554 |
+
"model.language_model.layers.1.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 555 |
+
"model.language_model.layers.1.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 556 |
+
"model.language_model.layers.1.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 557 |
+
"model.language_model.layers.10.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 558 |
+
"model.language_model.layers.10.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 559 |
+
"model.language_model.layers.10.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 560 |
+
"model.language_model.layers.10.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 561 |
+
"model.language_model.layers.10.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 562 |
+
"model.language_model.layers.10.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 563 |
+
"model.language_model.layers.10.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 564 |
+
"model.language_model.layers.10.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 565 |
+
"model.language_model.layers.10.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 566 |
+
"model.language_model.layers.10.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 567 |
+
"model.language_model.layers.10.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 568 |
+
"model.language_model.layers.11.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 569 |
+
"model.language_model.layers.11.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 570 |
+
"model.language_model.layers.11.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 571 |
+
"model.language_model.layers.11.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 572 |
+
"model.language_model.layers.11.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 573 |
+
"model.language_model.layers.11.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 574 |
+
"model.language_model.layers.11.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 575 |
+
"model.language_model.layers.11.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 576 |
+
"model.language_model.layers.11.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 577 |
+
"model.language_model.layers.11.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 578 |
+
"model.language_model.layers.11.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 579 |
+
"model.language_model.layers.12.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 580 |
+
"model.language_model.layers.12.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 581 |
+
"model.language_model.layers.12.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 582 |
+
"model.language_model.layers.12.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 583 |
+
"model.language_model.layers.12.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 584 |
+
"model.language_model.layers.12.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 585 |
+
"model.language_model.layers.12.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 586 |
+
"model.language_model.layers.12.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 587 |
+
"model.language_model.layers.12.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 588 |
+
"model.language_model.layers.12.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 589 |
+
"model.language_model.layers.12.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 590 |
+
"model.language_model.layers.13.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 591 |
+
"model.language_model.layers.13.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 592 |
+
"model.language_model.layers.13.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 593 |
+
"model.language_model.layers.13.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 594 |
+
"model.language_model.layers.13.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 595 |
+
"model.language_model.layers.13.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 596 |
+
"model.language_model.layers.13.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 597 |
+
"model.language_model.layers.13.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 598 |
+
"model.language_model.layers.13.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 599 |
+
"model.language_model.layers.13.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 600 |
+
"model.language_model.layers.13.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 601 |
+
"model.language_model.layers.14.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 602 |
+
"model.language_model.layers.14.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 603 |
+
"model.language_model.layers.14.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 604 |
+
"model.language_model.layers.14.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 605 |
+
"model.language_model.layers.14.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 606 |
+
"model.language_model.layers.14.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 607 |
+
"model.language_model.layers.14.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 608 |
+
"model.language_model.layers.14.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 609 |
+
"model.language_model.layers.14.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 610 |
+
"model.language_model.layers.14.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 611 |
+
"model.language_model.layers.14.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 612 |
+
"model.language_model.layers.15.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 613 |
+
"model.language_model.layers.15.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 614 |
+
"model.language_model.layers.15.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 615 |
+
"model.language_model.layers.15.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 616 |
+
"model.language_model.layers.15.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 617 |
+
"model.language_model.layers.15.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 618 |
+
"model.language_model.layers.15.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 619 |
+
"model.language_model.layers.15.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 620 |
+
"model.language_model.layers.15.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 621 |
+
"model.language_model.layers.15.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 622 |
+
"model.language_model.layers.15.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 623 |
+
"model.language_model.layers.16.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 624 |
+
"model.language_model.layers.16.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 625 |
+
"model.language_model.layers.16.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 626 |
+
"model.language_model.layers.16.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 627 |
+
"model.language_model.layers.16.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 628 |
+
"model.language_model.layers.16.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 629 |
+
"model.language_model.layers.16.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 630 |
+
"model.language_model.layers.16.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 631 |
+
"model.language_model.layers.16.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 632 |
+
"model.language_model.layers.16.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 633 |
+
"model.language_model.layers.16.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 634 |
+
"model.language_model.layers.17.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 635 |
+
"model.language_model.layers.17.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 636 |
+
"model.language_model.layers.17.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 637 |
+
"model.language_model.layers.17.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 638 |
+
"model.language_model.layers.17.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 639 |
+
"model.language_model.layers.17.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 640 |
+
"model.language_model.layers.17.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 641 |
+
"model.language_model.layers.17.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 642 |
+
"model.language_model.layers.17.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 643 |
+
"model.language_model.layers.17.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 644 |
+
"model.language_model.layers.17.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 645 |
+
"model.language_model.layers.18.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 646 |
+
"model.language_model.layers.18.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 647 |
+
"model.language_model.layers.18.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 648 |
+
"model.language_model.layers.18.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 649 |
+
"model.language_model.layers.18.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 650 |
+
"model.language_model.layers.18.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 651 |
+
"model.language_model.layers.18.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 652 |
+
"model.language_model.layers.18.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 653 |
+
"model.language_model.layers.18.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 654 |
+
"model.language_model.layers.18.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 655 |
+
"model.language_model.layers.18.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 656 |
+
"model.language_model.layers.19.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 657 |
+
"model.language_model.layers.19.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 658 |
+
"model.language_model.layers.19.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 659 |
+
"model.language_model.layers.19.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 660 |
+
"model.language_model.layers.19.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 661 |
+
"model.language_model.layers.19.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 662 |
+
"model.language_model.layers.19.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 663 |
+
"model.language_model.layers.19.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 664 |
+
"model.language_model.layers.19.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 665 |
+
"model.language_model.layers.19.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 666 |
+
"model.language_model.layers.19.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 667 |
+
"model.language_model.layers.2.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 668 |
+
"model.language_model.layers.2.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 669 |
+
"model.language_model.layers.2.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 670 |
+
"model.language_model.layers.2.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 671 |
+
"model.language_model.layers.2.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 672 |
+
"model.language_model.layers.2.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 673 |
+
"model.language_model.layers.2.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 674 |
+
"model.language_model.layers.2.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 675 |
+
"model.language_model.layers.2.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 676 |
+
"model.language_model.layers.2.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 677 |
+
"model.language_model.layers.2.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 678 |
+
"model.language_model.layers.20.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 679 |
+
"model.language_model.layers.20.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 680 |
+
"model.language_model.layers.20.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 681 |
+
"model.language_model.layers.20.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 682 |
+
"model.language_model.layers.20.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 683 |
+
"model.language_model.layers.20.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 684 |
+
"model.language_model.layers.20.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 685 |
+
"model.language_model.layers.20.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 686 |
+
"model.language_model.layers.20.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 687 |
+
"model.language_model.layers.20.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 688 |
+
"model.language_model.layers.20.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 689 |
+
"model.language_model.layers.21.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 690 |
+
"model.language_model.layers.21.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 691 |
+
"model.language_model.layers.21.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 692 |
+
"model.language_model.layers.21.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 693 |
+
"model.language_model.layers.21.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 694 |
+
"model.language_model.layers.21.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 695 |
+
"model.language_model.layers.21.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 696 |
+
"model.language_model.layers.21.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 697 |
+
"model.language_model.layers.21.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 698 |
+
"model.language_model.layers.21.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 699 |
+
"model.language_model.layers.21.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 700 |
+
"model.language_model.layers.22.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 701 |
+
"model.language_model.layers.22.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 702 |
+
"model.language_model.layers.22.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 703 |
+
"model.language_model.layers.22.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 704 |
+
"model.language_model.layers.22.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 705 |
+
"model.language_model.layers.22.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 706 |
+
"model.language_model.layers.22.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 707 |
+
"model.language_model.layers.22.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 708 |
+
"model.language_model.layers.22.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 709 |
+
"model.language_model.layers.22.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 710 |
+
"model.language_model.layers.22.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 711 |
+
"model.language_model.layers.23.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 712 |
+
"model.language_model.layers.23.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 713 |
+
"model.language_model.layers.23.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 714 |
+
"model.language_model.layers.23.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 715 |
+
"model.language_model.layers.23.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 716 |
+
"model.language_model.layers.23.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 717 |
+
"model.language_model.layers.23.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 718 |
+
"model.language_model.layers.23.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 719 |
+
"model.language_model.layers.23.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 720 |
+
"model.language_model.layers.23.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 721 |
+
"model.language_model.layers.23.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 722 |
+
"model.language_model.layers.24.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 723 |
+
"model.language_model.layers.24.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 724 |
+
"model.language_model.layers.24.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 725 |
+
"model.language_model.layers.24.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 726 |
+
"model.language_model.layers.24.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 727 |
+
"model.language_model.layers.24.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 728 |
+
"model.language_model.layers.24.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 729 |
+
"model.language_model.layers.24.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 730 |
+
"model.language_model.layers.24.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 731 |
+
"model.language_model.layers.24.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 732 |
+
"model.language_model.layers.24.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 733 |
+
"model.language_model.layers.25.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 734 |
+
"model.language_model.layers.25.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 735 |
+
"model.language_model.layers.25.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 736 |
+
"model.language_model.layers.25.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 737 |
+
"model.language_model.layers.25.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 738 |
+
"model.language_model.layers.25.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 739 |
+
"model.language_model.layers.25.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 740 |
+
"model.language_model.layers.25.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 741 |
+
"model.language_model.layers.25.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 742 |
+
"model.language_model.layers.25.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 743 |
+
"model.language_model.layers.25.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 744 |
+
"model.language_model.layers.26.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 745 |
+
"model.language_model.layers.26.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 746 |
+
"model.language_model.layers.26.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 747 |
+
"model.language_model.layers.26.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 748 |
+
"model.language_model.layers.26.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 749 |
+
"model.language_model.layers.26.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 750 |
+
"model.language_model.layers.26.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 751 |
+
"model.language_model.layers.26.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 752 |
+
"model.language_model.layers.26.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 753 |
+
"model.language_model.layers.26.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 754 |
+
"model.language_model.layers.26.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 755 |
+
"model.language_model.layers.27.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 756 |
+
"model.language_model.layers.27.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 757 |
+
"model.language_model.layers.27.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 758 |
+
"model.language_model.layers.27.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 759 |
+
"model.language_model.layers.27.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 760 |
+
"model.language_model.layers.27.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 761 |
+
"model.language_model.layers.27.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 762 |
+
"model.language_model.layers.27.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 763 |
+
"model.language_model.layers.27.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 764 |
+
"model.language_model.layers.27.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 765 |
+
"model.language_model.layers.27.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 766 |
+
"model.language_model.layers.3.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 767 |
+
"model.language_model.layers.3.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 768 |
+
"model.language_model.layers.3.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 769 |
+
"model.language_model.layers.3.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 770 |
+
"model.language_model.layers.3.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 771 |
+
"model.language_model.layers.3.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 772 |
+
"model.language_model.layers.3.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 773 |
+
"model.language_model.layers.3.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 774 |
+
"model.language_model.layers.3.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 775 |
+
"model.language_model.layers.3.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 776 |
+
"model.language_model.layers.3.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 777 |
+
"model.language_model.layers.4.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 778 |
+
"model.language_model.layers.4.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 779 |
+
"model.language_model.layers.4.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 780 |
+
"model.language_model.layers.4.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 781 |
+
"model.language_model.layers.4.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 782 |
+
"model.language_model.layers.4.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 783 |
+
"model.language_model.layers.4.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 784 |
+
"model.language_model.layers.4.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 785 |
+
"model.language_model.layers.4.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 786 |
+
"model.language_model.layers.4.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 787 |
+
"model.language_model.layers.4.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 788 |
+
"model.language_model.layers.5.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 789 |
+
"model.language_model.layers.5.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 790 |
+
"model.language_model.layers.5.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 791 |
+
"model.language_model.layers.5.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 792 |
+
"model.language_model.layers.5.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 793 |
+
"model.language_model.layers.5.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 794 |
+
"model.language_model.layers.5.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 795 |
+
"model.language_model.layers.5.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 796 |
+
"model.language_model.layers.5.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 797 |
+
"model.language_model.layers.5.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 798 |
+
"model.language_model.layers.5.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 799 |
+
"model.language_model.layers.6.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 800 |
+
"model.language_model.layers.6.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 801 |
+
"model.language_model.layers.6.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 802 |
+
"model.language_model.layers.6.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 803 |
+
"model.language_model.layers.6.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 804 |
+
"model.language_model.layers.6.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 805 |
+
"model.language_model.layers.6.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 806 |
+
"model.language_model.layers.6.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 807 |
+
"model.language_model.layers.6.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 808 |
+
"model.language_model.layers.6.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 809 |
+
"model.language_model.layers.6.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 810 |
+
"model.language_model.layers.7.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 811 |
+
"model.language_model.layers.7.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 812 |
+
"model.language_model.layers.7.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 813 |
+
"model.language_model.layers.7.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 814 |
+
"model.language_model.layers.7.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 815 |
+
"model.language_model.layers.7.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 816 |
+
"model.language_model.layers.7.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 817 |
+
"model.language_model.layers.7.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 818 |
+
"model.language_model.layers.7.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 819 |
+
"model.language_model.layers.7.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 820 |
+
"model.language_model.layers.7.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 821 |
+
"model.language_model.layers.8.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 822 |
+
"model.language_model.layers.8.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 823 |
+
"model.language_model.layers.8.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 824 |
+
"model.language_model.layers.8.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 825 |
+
"model.language_model.layers.8.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 826 |
+
"model.language_model.layers.8.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 827 |
+
"model.language_model.layers.8.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 828 |
+
"model.language_model.layers.8.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 829 |
+
"model.language_model.layers.8.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 830 |
+
"model.language_model.layers.8.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 831 |
+
"model.language_model.layers.8.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 832 |
+
"model.language_model.layers.9.post_attention_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 833 |
+
"model.language_model.layers.9.mlp.gate_proj.weight": "model-00000-of-00001.safetensors",
|
| 834 |
+
"model.language_model.layers.9.mlp.up_proj.weight": "model-00000-of-00001.safetensors",
|
| 835 |
+
"model.language_model.layers.9.mlp.down_proj.weight": "model-00000-of-00001.safetensors",
|
| 836 |
+
"model.language_model.layers.9.self_attn.k_norm.weight": "model-00000-of-00001.safetensors",
|
| 837 |
+
"model.language_model.layers.9.self_attn.o_proj.weight": "model-00000-of-00001.safetensors",
|
| 838 |
+
"model.language_model.layers.9.input_layernorm.weight": "model-00000-of-00001.safetensors",
|
| 839 |
+
"model.language_model.layers.9.self_attn.q_proj.weight": "model-00000-of-00001.safetensors",
|
| 840 |
+
"model.language_model.layers.9.self_attn.k_proj.weight": "model-00000-of-00001.safetensors",
|
| 841 |
+
"model.language_model.layers.9.self_attn.v_proj.weight": "model-00000-of-00001.safetensors",
|
| 842 |
+
"model.language_model.layers.9.self_attn.q_norm.weight": "model-00000-of-00001.safetensors",
|
| 843 |
+
"model.language_model.embed_tokens.weight": "model-00000-of-00001.safetensors"
|
| 844 |
+
}
|
| 845 |
+
}
|
modeling_Musci.py
ADDED
|
@@ -0,0 +1,333 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from typing import Optional, List, Union, Tuple
|
| 2 |
+
import torch
|
| 3 |
+
import torch.nn as nn
|
| 4 |
+
from transformers.modeling_outputs import CausalLMOutputWithPast, BaseModelOutputWithPast
|
| 5 |
+
from transformers.utils import logging
|
| 6 |
+
|
| 7 |
+
from transformers.models.qwen3.modeling_qwen3 import Qwen3Model, Qwen3PreTrainedModel, Qwen3DecoderLayer
|
| 8 |
+
from transformers.models.qwen3_omni_moe.modeling_qwen3_omni_moe import Qwen3OmniMoeAudioEncoder
|
| 9 |
+
from transformers.models.qwen3_omni_moe.configuration_qwen3_omni_moe import Qwen3OmniMoeAudioEncoderConfig
|
| 10 |
+
|
| 11 |
+
from transformers.models.qwen3.configuration_qwen3 import Qwen3Config
|
| 12 |
+
from transformers.utils.auto_docstring import auto_docstring
|
| 13 |
+
from transformers.modeling_utils import PreTrainedModel
|
| 14 |
+
from transformers.generation.utils import GenerationMixin
|
| 15 |
+
|
| 16 |
+
class MusciConfig(Qwen3Config):
|
| 17 |
+
model_type = "musci"
|
| 18 |
+
is_composition = True
|
| 19 |
+
# Make the architecture discoverable by Megatron-Bridge's AutoBridge
|
| 20 |
+
# when loading configs from disk.
|
| 21 |
+
architectures = ["MusciForCausalLM"]
|
| 22 |
+
|
| 23 |
+
def __init__(
|
| 24 |
+
self,
|
| 25 |
+
audio_config=None,
|
| 26 |
+
language_config=None,
|
| 27 |
+
adapter_hidden_size=8192,
|
| 28 |
+
ignore_index=-100,
|
| 29 |
+
**kwargs
|
| 30 |
+
):
|
| 31 |
+
num_hidden_layers = None
|
| 32 |
+
if language_config is not None:
|
| 33 |
+
if isinstance(language_config, dict):
|
| 34 |
+
num_hidden_layers = language_config.get("num_hidden_layers", None)
|
| 35 |
+
elif isinstance(language_config, Qwen3Config):
|
| 36 |
+
num_hidden_layers = language_config.num_hidden_layers
|
| 37 |
+
|
| 38 |
+
if num_hidden_layers is not None:
|
| 39 |
+
kwargs.update({"num_hidden_layers": num_hidden_layers})
|
| 40 |
+
|
| 41 |
+
# Initialize parent Qwen3Config with kwargs to handle standard config params
|
| 42 |
+
super().__init__(**kwargs)
|
| 43 |
+
|
| 44 |
+
if isinstance(audio_config, dict):
|
| 45 |
+
audio_config = Qwen3OmniMoeAudioEncoderConfig(**audio_config)
|
| 46 |
+
if isinstance(audio_config, Qwen3OmniMoeAudioEncoderConfig):
|
| 47 |
+
audio_config = audio_config
|
| 48 |
+
elif audio_config is None:
|
| 49 |
+
audio_config = Qwen3OmniMoeAudioEncoderConfig()
|
| 50 |
+
|
| 51 |
+
if isinstance(language_config, dict):
|
| 52 |
+
language_config = Qwen3Config(**language_config)
|
| 53 |
+
elif isinstance(language_config, Qwen3Config):
|
| 54 |
+
language_config = language_config
|
| 55 |
+
elif language_config is None:
|
| 56 |
+
language_config = Qwen3Config()
|
| 57 |
+
|
| 58 |
+
self.audio_config = audio_config
|
| 59 |
+
self.language_config = language_config
|
| 60 |
+
self.adapter_hidden_size = adapter_hidden_size
|
| 61 |
+
self.ignore_index = ignore_index
|
| 62 |
+
self.dtype = language_config.dtype
|
| 63 |
+
|
| 64 |
+
def to_dict(self):
|
| 65 |
+
output = super().to_dict()
|
| 66 |
+
if self.audio_config is not None:
|
| 67 |
+
if hasattr(self.audio_config, "to_dict"):
|
| 68 |
+
output["audio_config"] = self.audio_config.to_dict()
|
| 69 |
+
else:
|
| 70 |
+
output["audio_config"] = self.audio_config
|
| 71 |
+
if self.language_config is not None:
|
| 72 |
+
if hasattr(self.language_config, "to_dict"):
|
| 73 |
+
output["language_config"] = self.language_config.to_dict()
|
| 74 |
+
else:
|
| 75 |
+
output["language_config"] = self.language_config
|
| 76 |
+
return output
|
| 77 |
+
|
| 78 |
+
class MusciGatedMLP(nn.Module):
|
| 79 |
+
def __init__(self, input_size, hidden_size, output_size):
|
| 80 |
+
super().__init__()
|
| 81 |
+
self.gate_proj = nn.Linear(input_size, hidden_size, bias=False)
|
| 82 |
+
self.up_proj = nn.Linear(input_size, hidden_size, bias=False)
|
| 83 |
+
self.down_proj = nn.Linear(hidden_size, output_size, bias=False)
|
| 84 |
+
self.act_fn = nn.SiLU()
|
| 85 |
+
|
| 86 |
+
def forward(self, x):
|
| 87 |
+
return self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x))
|
| 88 |
+
|
| 89 |
+
@auto_docstring
|
| 90 |
+
class MusciPreTrainedModel(PreTrainedModel):
|
| 91 |
+
config: MusciConfig
|
| 92 |
+
base_model_prefix = "model"
|
| 93 |
+
supports_gradient_checkpointing = True
|
| 94 |
+
_no_split_modules = ["Qwen3DecoderLayer"]
|
| 95 |
+
_skip_keys_device_placement = ["past_key_values"]
|
| 96 |
+
_supports_flash_attn = True
|
| 97 |
+
_supports_sdpa = True
|
| 98 |
+
_supports_flex_attn = True
|
| 99 |
+
|
| 100 |
+
_can_compile_fullgraph = False
|
| 101 |
+
_supports_attention_backend = True
|
| 102 |
+
_can_record_outputs = {
|
| 103 |
+
"hidden_states": Qwen3DecoderLayer,
|
| 104 |
+
}
|
| 105 |
+
|
| 106 |
+
class MusciModel(MusciPreTrainedModel):
|
| 107 |
+
config_class = MusciConfig
|
| 108 |
+
|
| 109 |
+
def __init__(self, config: MusciConfig):
|
| 110 |
+
super().__init__(config)
|
| 111 |
+
|
| 112 |
+
self.audio_model = Qwen3OmniMoeAudioEncoder(config.audio_config)
|
| 113 |
+
self.language_model = Qwen3Model(config.language_config)
|
| 114 |
+
|
| 115 |
+
self.audio_adapter = MusciGatedMLP(
|
| 116 |
+
input_size=config.audio_config.output_dim,
|
| 117 |
+
hidden_size=config.adapter_hidden_size,
|
| 118 |
+
output_size=config.language_config.hidden_size
|
| 119 |
+
)
|
| 120 |
+
|
| 121 |
+
# Initialize weights and apply final processing
|
| 122 |
+
self.post_init()
|
| 123 |
+
|
| 124 |
+
def get_input_embeddings(self):
|
| 125 |
+
return self.language_model.get_input_embeddings()
|
| 126 |
+
|
| 127 |
+
def set_input_embeddings(self, value):
|
| 128 |
+
self.language_model.set_input_embeddings(value)
|
| 129 |
+
|
| 130 |
+
def get_audio_features(self, input_features, feature_lens):
|
| 131 |
+
audio_outputs = self.audio_model(
|
| 132 |
+
input_features=input_features,
|
| 133 |
+
feature_lens=feature_lens,
|
| 134 |
+
)
|
| 135 |
+
return audio_outputs.last_hidden_state
|
| 136 |
+
|
| 137 |
+
def forward(
|
| 138 |
+
self,
|
| 139 |
+
input_ids: torch.LongTensor = None,
|
| 140 |
+
attention_mask: Optional[torch.Tensor] = None,
|
| 141 |
+
position_ids: Optional[torch.LongTensor] = None,
|
| 142 |
+
past_key_values: Optional[List[torch.FloatTensor]] = None,
|
| 143 |
+
inputs_embeds: Optional[torch.FloatTensor] = None,
|
| 144 |
+
use_cache: Optional[bool] = None,
|
| 145 |
+
output_attentions: Optional[bool] = None,
|
| 146 |
+
output_hidden_states: Optional[bool] = None,
|
| 147 |
+
return_dict: Optional[bool] = None,
|
| 148 |
+
audio_data: Optional[torch.FloatTensor] = None,
|
| 149 |
+
audio_data_seqlens: Optional[torch.Tensor] = None,
|
| 150 |
+
audio_input_mask: Optional[torch.Tensor] = None,
|
| 151 |
+
cache_position: Optional[torch.LongTensor] = None,
|
| 152 |
+
) -> Union[Tuple, BaseModelOutputWithPast]:
|
| 153 |
+
output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
|
| 154 |
+
output_hidden_states = (
|
| 155 |
+
output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
|
| 156 |
+
)
|
| 157 |
+
return_dict = return_dict if return_dict is not None else self.config.use_return_dict
|
| 158 |
+
|
| 159 |
+
# 1. Get text embeddings
|
| 160 |
+
if inputs_embeds is None:
|
| 161 |
+
inputs_embeds = self.get_input_embeddings()(input_ids)
|
| 162 |
+
|
| 163 |
+
# 2. Process audio and merge embeddings if audio is present
|
| 164 |
+
if audio_data is not None:
|
| 165 |
+
# [B, Audio_Len, D]
|
| 166 |
+
audio_embeds = self.get_audio_features(audio_data, audio_data_seqlens)
|
| 167 |
+
audio_embeds = self.audio_adapter(audio_embeds)
|
| 168 |
+
|
| 169 |
+
# audio_input_mask: [B, L] -> [B, L, 1] -> [B, L, D]
|
| 170 |
+
# D elements will be replaced by audio embeddings
|
| 171 |
+
mask_expanded = audio_input_mask.unsqueeze(-1).expand_as(inputs_embeds)
|
| 172 |
+
inputs_embeds.masked_scatter_(mask_expanded, audio_embeds)
|
| 173 |
+
|
| 174 |
+
# 3. Forward pass through language model
|
| 175 |
+
return self.language_model(
|
| 176 |
+
input_ids=None, # We pass inputs_embeds
|
| 177 |
+
attention_mask=attention_mask,
|
| 178 |
+
position_ids=position_ids,
|
| 179 |
+
past_key_values=past_key_values,
|
| 180 |
+
inputs_embeds=inputs_embeds,
|
| 181 |
+
use_cache=use_cache,
|
| 182 |
+
output_attentions=output_attentions,
|
| 183 |
+
output_hidden_states=output_hidden_states,
|
| 184 |
+
return_dict=return_dict,
|
| 185 |
+
cache_position=cache_position,
|
| 186 |
+
)
|
| 187 |
+
|
| 188 |
+
class MusciForCausalLM(MusciPreTrainedModel, GenerationMixin):
|
| 189 |
+
config_class = MusciConfig
|
| 190 |
+
_tied_weights_keys = ["lm_head.weight"]
|
| 191 |
+
_keys_to_ignore_on_save = ["lm_head.weight"]
|
| 192 |
+
|
| 193 |
+
def __init__(self, config: MusciConfig):
|
| 194 |
+
super().__init__(config)
|
| 195 |
+
self.model = MusciModel(config)
|
| 196 |
+
self.vocab_size = config.language_config.vocab_size
|
| 197 |
+
self.lm_head = nn.Linear(config.language_config.hidden_size, self.vocab_size, bias=False)
|
| 198 |
+
|
| 199 |
+
# Initialize weights and apply final processing
|
| 200 |
+
self.post_init()
|
| 201 |
+
|
| 202 |
+
def tie_weights(self):
|
| 203 |
+
super().tie_weights()
|
| 204 |
+
|
| 205 |
+
# tie lm_head to input embeddings
|
| 206 |
+
self.lm_head.weight = self.model.language_model.embed_tokens.weight
|
| 207 |
+
|
| 208 |
+
def get_input_embeddings(self):
|
| 209 |
+
return self.model.get_input_embeddings()
|
| 210 |
+
|
| 211 |
+
def set_input_embeddings(self, value):
|
| 212 |
+
self.model.set_input_embeddings(value)
|
| 213 |
+
|
| 214 |
+
def get_output_embeddings(self):
|
| 215 |
+
return self.lm_head
|
| 216 |
+
|
| 217 |
+
def set_output_embeddings(self, new_embeddings):
|
| 218 |
+
self.lm_head = new_embeddings
|
| 219 |
+
|
| 220 |
+
def forward(
|
| 221 |
+
self,
|
| 222 |
+
input_ids: torch.LongTensor = None,
|
| 223 |
+
attention_mask: Optional[torch.Tensor] = None,
|
| 224 |
+
position_ids: Optional[torch.LongTensor] = None,
|
| 225 |
+
past_key_values: Optional[List[torch.FloatTensor]] = None,
|
| 226 |
+
inputs_embeds: Optional[torch.FloatTensor] = None,
|
| 227 |
+
labels: Optional[torch.LongTensor] = None,
|
| 228 |
+
use_cache: Optional[bool] = None,
|
| 229 |
+
output_attentions: Optional[bool] = None,
|
| 230 |
+
output_hidden_states: Optional[bool] = None,
|
| 231 |
+
return_dict: Optional[bool] = None,
|
| 232 |
+
audio_data: Optional[torch.FloatTensor] = None,
|
| 233 |
+
audio_data_seqlens: Optional[torch.Tensor] = None,
|
| 234 |
+
audio_input_mask: Optional[torch.Tensor] = None,
|
| 235 |
+
cache_position: Optional[torch.LongTensor] = None,
|
| 236 |
+
) -> Union[Tuple, CausalLMOutputWithPast]:
|
| 237 |
+
output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
|
| 238 |
+
output_hidden_states = (
|
| 239 |
+
output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
|
| 240 |
+
)
|
| 241 |
+
return_dict = return_dict if return_dict is not None else self.config.use_return_dict
|
| 242 |
+
|
| 243 |
+
outputs = self.model(
|
| 244 |
+
input_ids=input_ids,
|
| 245 |
+
attention_mask=attention_mask,
|
| 246 |
+
position_ids=position_ids,
|
| 247 |
+
past_key_values=past_key_values,
|
| 248 |
+
inputs_embeds=inputs_embeds,
|
| 249 |
+
use_cache=use_cache,
|
| 250 |
+
output_attentions=output_attentions,
|
| 251 |
+
output_hidden_states=output_hidden_states,
|
| 252 |
+
return_dict=return_dict,
|
| 253 |
+
audio_data=audio_data,
|
| 254 |
+
audio_data_seqlens=audio_data_seqlens,
|
| 255 |
+
audio_input_mask=audio_input_mask,
|
| 256 |
+
cache_position=cache_position,
|
| 257 |
+
)
|
| 258 |
+
|
| 259 |
+
hidden_states = outputs[0]
|
| 260 |
+
logits = self.lm_head(hidden_states)
|
| 261 |
+
|
| 262 |
+
loss = None
|
| 263 |
+
if labels is not None:
|
| 264 |
+
# Shift so that tokens < n predict n
|
| 265 |
+
shift_logits = logits[..., :-1, :].contiguous()
|
| 266 |
+
shift_labels = labels[..., 1:].contiguous()
|
| 267 |
+
# Flatten the tokens
|
| 268 |
+
loss_fct = nn.CrossEntropyLoss(ignore_index=self.config.ignore_index)
|
| 269 |
+
shift_logits = shift_logits.view(-1, self.config.language_config.vocab_size)
|
| 270 |
+
shift_labels = shift_labels.view(-1)
|
| 271 |
+
# Enable model parallelism
|
| 272 |
+
shift_labels = shift_labels.to(shift_logits.device)
|
| 273 |
+
loss = loss_fct(shift_logits, shift_labels)
|
| 274 |
+
|
| 275 |
+
if not return_dict:
|
| 276 |
+
output = (logits,) + outputs[1:]
|
| 277 |
+
return ((loss,) + output) if loss is not None else output
|
| 278 |
+
|
| 279 |
+
return CausalLMOutputWithPast(
|
| 280 |
+
loss=loss,
|
| 281 |
+
logits=logits,
|
| 282 |
+
past_key_values=outputs.past_key_values,
|
| 283 |
+
hidden_states=outputs.hidden_states,
|
| 284 |
+
attentions=outputs.attentions,
|
| 285 |
+
)
|
| 286 |
+
|
| 287 |
+
def prepare_inputs_for_generation(
|
| 288 |
+
self,
|
| 289 |
+
input_ids,
|
| 290 |
+
past_key_values=None,
|
| 291 |
+
attention_mask=None,
|
| 292 |
+
inputs_embeds=None,
|
| 293 |
+
cache_position=None,
|
| 294 |
+
**kwargs
|
| 295 |
+
):
|
| 296 |
+
# decoding step (KV cache present) keeps only the last token and drops audio inputs;
|
| 297 |
+
# prefill step pulls audio inputs from kwargs.
|
| 298 |
+
position_ids = kwargs.get("position_ids", None)
|
| 299 |
+
if cache_position is not None and cache_position[0] > 0:
|
| 300 |
+
input_ids = input_ids[:, -1:]
|
| 301 |
+
if position_ids is not None:
|
| 302 |
+
position_ids = position_ids[:, -1:]
|
| 303 |
+
audio_data = None
|
| 304 |
+
audio_input_mask = None
|
| 305 |
+
audio_data_seqlens = None
|
| 306 |
+
else:
|
| 307 |
+
audio_data = kwargs.get("audio_data", None)
|
| 308 |
+
audio_input_mask = kwargs.get("audio_input_mask", None)
|
| 309 |
+
audio_data_seqlens = kwargs.get("audio_data_seqlens", None)
|
| 310 |
+
|
| 311 |
+
# prefer inputs_embeds at the first step when present
|
| 312 |
+
if inputs_embeds is not None and past_key_values is None:
|
| 313 |
+
model_inputs = {"inputs_embeds": inputs_embeds}
|
| 314 |
+
else:
|
| 315 |
+
model_inputs = {"input_ids": input_ids}
|
| 316 |
+
|
| 317 |
+
model_inputs.update({
|
| 318 |
+
"past_key_values": past_key_values,
|
| 319 |
+
"use_cache": kwargs.get("use_cache"),
|
| 320 |
+
"attention_mask": attention_mask,
|
| 321 |
+
"position_ids": position_ids,
|
| 322 |
+
"audio_data": audio_data,
|
| 323 |
+
"audio_input_mask": audio_input_mask,
|
| 324 |
+
"audio_data_seqlens": audio_data_seqlens,
|
| 325 |
+
})
|
| 326 |
+
|
| 327 |
+
return model_inputs
|
| 328 |
+
|
| 329 |
+
__all__ = [
|
| 330 |
+
"MusciConfig",
|
| 331 |
+
"MusciModel",
|
| 332 |
+
"MusciForCausalLM",
|
| 333 |
+
]
|
processing_Musci.py
ADDED
|
@@ -0,0 +1,219 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from dataclasses import dataclass
|
| 2 |
+
from typing import Optional, Union, List
|
| 3 |
+
import torch
|
| 4 |
+
import numpy as np
|
| 5 |
+
from transformers import BatchEncoding
|
| 6 |
+
from transformers.models.whisper.feature_extraction_whisper import WhisperFeatureExtractor
|
| 7 |
+
import importlib.util
|
| 8 |
+
import sys
|
| 9 |
+
|
| 10 |
+
|
| 11 |
+
@dataclass
|
| 12 |
+
class MelConfig:
|
| 13 |
+
mel_sr: int = 16000
|
| 14 |
+
mel_dim: int = 80
|
| 15 |
+
mel_n_fft: int = 640
|
| 16 |
+
mel_hop_length: int = 160
|
| 17 |
+
mel_dtype: torch.dtype = torch.bfloat16
|
| 18 |
+
|
| 19 |
+
|
| 20 |
+
def load_chat_template(template_path: str, package_path: Optional[str] = None) -> List:
|
| 21 |
+
"""Dynamically import a chat template module by file path and return its `chat_template`."""
|
| 22 |
+
import os
|
| 23 |
+
|
| 24 |
+
if package_path and package_path not in sys.path:
|
| 25 |
+
sys.path.insert(0, package_path)
|
| 26 |
+
|
| 27 |
+
spec = importlib.util.spec_from_file_location("chat_template_module", template_path)
|
| 28 |
+
module = importlib.util.module_from_spec(spec)
|
| 29 |
+
sys.modules["chat_template_module"] = module
|
| 30 |
+
spec.loader.exec_module(module)
|
| 31 |
+
return module.chat_template
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
class MusciProcessor:
|
| 35 |
+
"""Audio processor for Musci ASR: mel-spectrogram + chat-template-driven token layout."""
|
| 36 |
+
|
| 37 |
+
def __init__(
|
| 38 |
+
self,
|
| 39 |
+
tokenizer,
|
| 40 |
+
config: Optional[MelConfig] = None,
|
| 41 |
+
template_path: Optional[str] = None,
|
| 42 |
+
enable_time_marker: bool = False,
|
| 43 |
+
):
|
| 44 |
+
self.tokenizer = tokenizer
|
| 45 |
+
self.config = config or MelConfig()
|
| 46 |
+
|
| 47 |
+
# Whisper log-mel frontend — matches the front-end the model was trained with.
|
| 48 |
+
self.feature_extractor = WhisperFeatureExtractor(
|
| 49 |
+
feature_size=int(self.config.mel_dim),
|
| 50 |
+
sampling_rate=int(self.config.mel_sr),
|
| 51 |
+
hop_length=int(self.config.mel_hop_length),
|
| 52 |
+
n_fft=int(self.config.mel_n_fft),
|
| 53 |
+
)
|
| 54 |
+
|
| 55 |
+
# Special token ids (Qwen3 tokenizer).
|
| 56 |
+
self.start_token_id = 151644
|
| 57 |
+
self.end_token_id = 151645
|
| 58 |
+
self.audio_start_token_id = 151669
|
| 59 |
+
self.audio_end_token_id = 151670
|
| 60 |
+
self.audio_placeholder_id = 0
|
| 61 |
+
|
| 62 |
+
self.chat_template = None if template_path is None else load_chat_template(template_path)
|
| 63 |
+
self.enable_time_marker = enable_time_marker
|
| 64 |
+
|
| 65 |
+
# Digit tokens 0-9 in the Qwen3 tokenizer, used for time markers.
|
| 66 |
+
self._digit_token_ids = {str(d): 15 + d for d in range(10)}
|
| 67 |
+
self.audio_tokens_per_second = 12.5
|
| 68 |
+
self.time_marker_every_seconds = 2
|
| 69 |
+
self.time_marker_every_audio_tokens = int(
|
| 70 |
+
self.audio_tokens_per_second * self.time_marker_every_seconds
|
| 71 |
+
)
|
| 72 |
+
|
| 73 |
+
def load_template(self, template_path: str):
|
| 74 |
+
self.chat_template = load_chat_template(template_path)
|
| 75 |
+
print(f"Loaded chat template from {template_path}")
|
| 76 |
+
return self
|
| 77 |
+
|
| 78 |
+
def _get_feat_extract_output_lengths(self, input_lengths):
|
| 79 |
+
"""Map raw mel-frame count to number of audio tokens after the encoder downsample."""
|
| 80 |
+
input_lengths_leave = input_lengths % 100
|
| 81 |
+
feat_lengths = (input_lengths_leave - 1) // 2 + 1
|
| 82 |
+
output_lengths = (
|
| 83 |
+
((feat_lengths - 1) // 2 + 1 - 1) // 2 + 1 + (input_lengths // 100) * 13
|
| 84 |
+
)
|
| 85 |
+
return output_lengths
|
| 86 |
+
|
| 87 |
+
def _get_time_marker_token_ids(self, second: int) -> List[int]:
|
| 88 |
+
return [self._digit_token_ids[c] for c in str(second)]
|
| 89 |
+
|
| 90 |
+
def _build_audio_tokens_with_time_markers(self, audio_seq_len: int) -> List[int]:
|
| 91 |
+
"""Interleave time markers every `time_marker_every_seconds` seconds of audio tokens."""
|
| 92 |
+
num_full_seconds = int(audio_seq_len / self.audio_tokens_per_second)
|
| 93 |
+
|
| 94 |
+
tokens_list: List[int] = []
|
| 95 |
+
audio_tokens_consumed = 0
|
| 96 |
+
|
| 97 |
+
for second in range(
|
| 98 |
+
self.time_marker_every_seconds, num_full_seconds + 1, self.time_marker_every_seconds
|
| 99 |
+
):
|
| 100 |
+
marker_pos = (
|
| 101 |
+
(second // self.time_marker_every_seconds) * self.time_marker_every_audio_tokens
|
| 102 |
+
)
|
| 103 |
+
segment_len = marker_pos - audio_tokens_consumed
|
| 104 |
+
if segment_len > 0:
|
| 105 |
+
tokens_list.extend([self.audio_placeholder_id] * segment_len)
|
| 106 |
+
audio_tokens_consumed += segment_len
|
| 107 |
+
tokens_list.extend(self._get_time_marker_token_ids(second))
|
| 108 |
+
|
| 109 |
+
remaining = audio_seq_len - audio_tokens_consumed
|
| 110 |
+
if remaining > 0:
|
| 111 |
+
tokens_list.extend([self.audio_placeholder_id] * remaining)
|
| 112 |
+
return tokens_list
|
| 113 |
+
|
| 114 |
+
def _build_input_from_template(self, num_audio_tokens: int) -> tuple:
|
| 115 |
+
"""Walk the loaded chat_template and emit (input_ids, audio_input_mask) for inference."""
|
| 116 |
+
if self.chat_template is None:
|
| 117 |
+
raise ValueError("Chat template not loaded. Call load_template() first.")
|
| 118 |
+
|
| 119 |
+
input_ids: List[int] = []
|
| 120 |
+
audio_mask: List[bool] = []
|
| 121 |
+
|
| 122 |
+
for segment in self.chat_template:
|
| 123 |
+
seg_type = segment.type
|
| 124 |
+
|
| 125 |
+
if seg_type == "constant_text_token":
|
| 126 |
+
text_ids = segment.text_ids.tolist()
|
| 127 |
+
input_ids.extend(text_ids)
|
| 128 |
+
audio_mask.extend([False] * len(text_ids))
|
| 129 |
+
|
| 130 |
+
elif seg_type in ("audio_contiguous", "audio_token"):
|
| 131 |
+
if self.enable_time_marker:
|
| 132 |
+
audio_ids = self._build_audio_tokens_with_time_markers(num_audio_tokens)
|
| 133 |
+
input_ids.extend(audio_ids)
|
| 134 |
+
audio_mask.extend(
|
| 135 |
+
[tok == self.audio_placeholder_id for tok in audio_ids]
|
| 136 |
+
)
|
| 137 |
+
else:
|
| 138 |
+
input_ids.extend([self.audio_placeholder_id] * num_audio_tokens)
|
| 139 |
+
audio_mask.extend([True] * num_audio_tokens)
|
| 140 |
+
|
| 141 |
+
elif seg_type == "text_token":
|
| 142 |
+
# Generation starts here at inference time.
|
| 143 |
+
break
|
| 144 |
+
|
| 145 |
+
return input_ids, audio_mask
|
| 146 |
+
|
| 147 |
+
def _build_input_legacy(self, num_audio_tokens: int) -> tuple:
|
| 148 |
+
"""Hardcoded [start, audio_start, audio*, audio_end] layout, used when no template is loaded."""
|
| 149 |
+
if self.enable_time_marker:
|
| 150 |
+
audio_ids = self._build_audio_tokens_with_time_markers(num_audio_tokens)
|
| 151 |
+
ids = (
|
| 152 |
+
[self.start_token_id, self.audio_start_token_id]
|
| 153 |
+
+ audio_ids
|
| 154 |
+
+ [self.audio_end_token_id]
|
| 155 |
+
)
|
| 156 |
+
audio_mask = [tok == self.audio_placeholder_id for tok in audio_ids]
|
| 157 |
+
mask = [False, False] + audio_mask + [False]
|
| 158 |
+
else:
|
| 159 |
+
ids = (
|
| 160 |
+
[self.start_token_id, self.audio_start_token_id]
|
| 161 |
+
+ [self.audio_placeholder_id] * num_audio_tokens
|
| 162 |
+
+ [self.audio_end_token_id]
|
| 163 |
+
)
|
| 164 |
+
mask = [False, False] + [True] * num_audio_tokens + [False]
|
| 165 |
+
return ids, mask
|
| 166 |
+
|
| 167 |
+
def __call__(
|
| 168 |
+
self,
|
| 169 |
+
audio: Union[np.ndarray, torch.Tensor],
|
| 170 |
+
return_tensors: str = "pt",
|
| 171 |
+
**kwargs,
|
| 172 |
+
):
|
| 173 |
+
if audio is None:
|
| 174 |
+
raise ValueError("Audio input is required.")
|
| 175 |
+
|
| 176 |
+
if isinstance(audio, torch.Tensor):
|
| 177 |
+
waveform = audio.detach().to(dtype=torch.float32).cpu().numpy()
|
| 178 |
+
else:
|
| 179 |
+
waveform = np.asarray(audio, dtype=np.float32)
|
| 180 |
+
if waveform.ndim == 2:
|
| 181 |
+
waveform = waveform[0]
|
| 182 |
+
|
| 183 |
+
try:
|
| 184 |
+
mel = self.feature_extractor._np_extract_fbank_features(
|
| 185 |
+
waveform[None, ...], device="cpu"
|
| 186 |
+
)[0]
|
| 187 |
+
except TypeError:
|
| 188 |
+
mel = self.feature_extractor._np_extract_fbank_features(waveform[None, ...])[0]
|
| 189 |
+
input_features = torch.from_numpy(mel).to(self.config.mel_dtype)
|
| 190 |
+
if input_features.dim() == 3:
|
| 191 |
+
input_features = input_features.squeeze(0)
|
| 192 |
+
|
| 193 |
+
raw_mel_len = input_features.shape[-1]
|
| 194 |
+
num_audio_tokens = self._get_feat_extract_output_lengths(raw_mel_len)
|
| 195 |
+
|
| 196 |
+
if self.chat_template is not None:
|
| 197 |
+
ids, mask = self._build_input_from_template(num_audio_tokens)
|
| 198 |
+
else:
|
| 199 |
+
ids, mask = self._build_input_legacy(num_audio_tokens)
|
| 200 |
+
|
| 201 |
+
input_ids_tensor = torch.tensor([ids], dtype=torch.long)
|
| 202 |
+
audio_mask_tensor = torch.tensor([mask], dtype=torch.bool)
|
| 203 |
+
attention_mask_tensor = torch.ones_like(input_ids_tensor)
|
| 204 |
+
seq_lens_tensor = torch.tensor([raw_mel_len], dtype=torch.long)
|
| 205 |
+
|
| 206 |
+
data = {
|
| 207 |
+
"input_ids": input_ids_tensor,
|
| 208 |
+
"attention_mask": attention_mask_tensor,
|
| 209 |
+
"audio_data": input_features,
|
| 210 |
+
"audio_data_seqlens": seq_lens_tensor,
|
| 211 |
+
"audio_input_mask": audio_mask_tensor,
|
| 212 |
+
}
|
| 213 |
+
return BatchEncoding(data=data, tensor_type=return_tensors)
|
| 214 |
+
|
| 215 |
+
def batch_decode(self, *args, **kwargs):
|
| 216 |
+
return self.tokenizer.batch_decode(*args, **kwargs)
|
| 217 |
+
|
| 218 |
+
def decode(self, *args, **kwargs):
|
| 219 |
+
return self.tokenizer.decode(*args, **kwargs)
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<|im_start|>",
|
| 4 |
+
"<|im_end|>",
|
| 5 |
+
"<|object_ref_start|>",
|
| 6 |
+
"<|object_ref_end|>",
|
| 7 |
+
"<|box_start|>",
|
| 8 |
+
"<|box_end|>",
|
| 9 |
+
"<|quad_start|>",
|
| 10 |
+
"<|quad_end|>",
|
| 11 |
+
"<|vision_start|>",
|
| 12 |
+
"<|vision_end|>",
|
| 13 |
+
"<|vision_pad|>",
|
| 14 |
+
"<|image_pad|>",
|
| 15 |
+
"<|video_pad|>"
|
| 16 |
+
],
|
| 17 |
+
"eos_token": {
|
| 18 |
+
"content": "<|im_end|>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
},
|
| 24 |
+
"pad_token": {
|
| 25 |
+
"content": "<|endoftext|>",
|
| 26 |
+
"lstrip": false,
|
| 27 |
+
"normalized": false,
|
| 28 |
+
"rstrip": false,
|
| 29 |
+
"single_word": false
|
| 30 |
+
}
|
| 31 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
|
| 3 |
+
size 11422654
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,239 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": false,
|
| 3 |
+
"add_prefix_space": false,
|
| 4 |
+
"added_tokens_decoder": {
|
| 5 |
+
"151643": {
|
| 6 |
+
"content": "<|endoftext|>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false,
|
| 11 |
+
"special": true
|
| 12 |
+
},
|
| 13 |
+
"151644": {
|
| 14 |
+
"content": "<|im_start|>",
|
| 15 |
+
"lstrip": false,
|
| 16 |
+
"normalized": false,
|
| 17 |
+
"rstrip": false,
|
| 18 |
+
"single_word": false,
|
| 19 |
+
"special": true
|
| 20 |
+
},
|
| 21 |
+
"151645": {
|
| 22 |
+
"content": "<|im_end|>",
|
| 23 |
+
"lstrip": false,
|
| 24 |
+
"normalized": false,
|
| 25 |
+
"rstrip": false,
|
| 26 |
+
"single_word": false,
|
| 27 |
+
"special": true
|
| 28 |
+
},
|
| 29 |
+
"151646": {
|
| 30 |
+
"content": "<|object_ref_start|>",
|
| 31 |
+
"lstrip": false,
|
| 32 |
+
"normalized": false,
|
| 33 |
+
"rstrip": false,
|
| 34 |
+
"single_word": false,
|
| 35 |
+
"special": true
|
| 36 |
+
},
|
| 37 |
+
"151647": {
|
| 38 |
+
"content": "<|object_ref_end|>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false,
|
| 43 |
+
"special": true
|
| 44 |
+
},
|
| 45 |
+
"151648": {
|
| 46 |
+
"content": "<|box_start|>",
|
| 47 |
+
"lstrip": false,
|
| 48 |
+
"normalized": false,
|
| 49 |
+
"rstrip": false,
|
| 50 |
+
"single_word": false,
|
| 51 |
+
"special": true
|
| 52 |
+
},
|
| 53 |
+
"151649": {
|
| 54 |
+
"content": "<|box_end|>",
|
| 55 |
+
"lstrip": false,
|
| 56 |
+
"normalized": false,
|
| 57 |
+
"rstrip": false,
|
| 58 |
+
"single_word": false,
|
| 59 |
+
"special": true
|
| 60 |
+
},
|
| 61 |
+
"151650": {
|
| 62 |
+
"content": "<|quad_start|>",
|
| 63 |
+
"lstrip": false,
|
| 64 |
+
"normalized": false,
|
| 65 |
+
"rstrip": false,
|
| 66 |
+
"single_word": false,
|
| 67 |
+
"special": true
|
| 68 |
+
},
|
| 69 |
+
"151651": {
|
| 70 |
+
"content": "<|quad_end|>",
|
| 71 |
+
"lstrip": false,
|
| 72 |
+
"normalized": false,
|
| 73 |
+
"rstrip": false,
|
| 74 |
+
"single_word": false,
|
| 75 |
+
"special": true
|
| 76 |
+
},
|
| 77 |
+
"151652": {
|
| 78 |
+
"content": "<|vision_start|>",
|
| 79 |
+
"lstrip": false,
|
| 80 |
+
"normalized": false,
|
| 81 |
+
"rstrip": false,
|
| 82 |
+
"single_word": false,
|
| 83 |
+
"special": true
|
| 84 |
+
},
|
| 85 |
+
"151653": {
|
| 86 |
+
"content": "<|vision_end|>",
|
| 87 |
+
"lstrip": false,
|
| 88 |
+
"normalized": false,
|
| 89 |
+
"rstrip": false,
|
| 90 |
+
"single_word": false,
|
| 91 |
+
"special": true
|
| 92 |
+
},
|
| 93 |
+
"151654": {
|
| 94 |
+
"content": "<|vision_pad|>",
|
| 95 |
+
"lstrip": false,
|
| 96 |
+
"normalized": false,
|
| 97 |
+
"rstrip": false,
|
| 98 |
+
"single_word": false,
|
| 99 |
+
"special": true
|
| 100 |
+
},
|
| 101 |
+
"151655": {
|
| 102 |
+
"content": "<|image_pad|>",
|
| 103 |
+
"lstrip": false,
|
| 104 |
+
"normalized": false,
|
| 105 |
+
"rstrip": false,
|
| 106 |
+
"single_word": false,
|
| 107 |
+
"special": true
|
| 108 |
+
},
|
| 109 |
+
"151656": {
|
| 110 |
+
"content": "<|video_pad|>",
|
| 111 |
+
"lstrip": false,
|
| 112 |
+
"normalized": false,
|
| 113 |
+
"rstrip": false,
|
| 114 |
+
"single_word": false,
|
| 115 |
+
"special": true
|
| 116 |
+
},
|
| 117 |
+
"151657": {
|
| 118 |
+
"content": "<tool_call>",
|
| 119 |
+
"lstrip": false,
|
| 120 |
+
"normalized": false,
|
| 121 |
+
"rstrip": false,
|
| 122 |
+
"single_word": false,
|
| 123 |
+
"special": false
|
| 124 |
+
},
|
| 125 |
+
"151658": {
|
| 126 |
+
"content": "</tool_call>",
|
| 127 |
+
"lstrip": false,
|
| 128 |
+
"normalized": false,
|
| 129 |
+
"rstrip": false,
|
| 130 |
+
"single_word": false,
|
| 131 |
+
"special": false
|
| 132 |
+
},
|
| 133 |
+
"151659": {
|
| 134 |
+
"content": "<|fim_prefix|>",
|
| 135 |
+
"lstrip": false,
|
| 136 |
+
"normalized": false,
|
| 137 |
+
"rstrip": false,
|
| 138 |
+
"single_word": false,
|
| 139 |
+
"special": false
|
| 140 |
+
},
|
| 141 |
+
"151660": {
|
| 142 |
+
"content": "<|fim_middle|>",
|
| 143 |
+
"lstrip": false,
|
| 144 |
+
"normalized": false,
|
| 145 |
+
"rstrip": false,
|
| 146 |
+
"single_word": false,
|
| 147 |
+
"special": false
|
| 148 |
+
},
|
| 149 |
+
"151661": {
|
| 150 |
+
"content": "<|fim_suffix|>",
|
| 151 |
+
"lstrip": false,
|
| 152 |
+
"normalized": false,
|
| 153 |
+
"rstrip": false,
|
| 154 |
+
"single_word": false,
|
| 155 |
+
"special": false
|
| 156 |
+
},
|
| 157 |
+
"151662": {
|
| 158 |
+
"content": "<|fim_pad|>",
|
| 159 |
+
"lstrip": false,
|
| 160 |
+
"normalized": false,
|
| 161 |
+
"rstrip": false,
|
| 162 |
+
"single_word": false,
|
| 163 |
+
"special": false
|
| 164 |
+
},
|
| 165 |
+
"151663": {
|
| 166 |
+
"content": "<|repo_name|>",
|
| 167 |
+
"lstrip": false,
|
| 168 |
+
"normalized": false,
|
| 169 |
+
"rstrip": false,
|
| 170 |
+
"single_word": false,
|
| 171 |
+
"special": false
|
| 172 |
+
},
|
| 173 |
+
"151664": {
|
| 174 |
+
"content": "<|file_sep|>",
|
| 175 |
+
"lstrip": false,
|
| 176 |
+
"normalized": false,
|
| 177 |
+
"rstrip": false,
|
| 178 |
+
"single_word": false,
|
| 179 |
+
"special": false
|
| 180 |
+
},
|
| 181 |
+
"151665": {
|
| 182 |
+
"content": "<tool_response>",
|
| 183 |
+
"lstrip": false,
|
| 184 |
+
"normalized": false,
|
| 185 |
+
"rstrip": false,
|
| 186 |
+
"single_word": false,
|
| 187 |
+
"special": false
|
| 188 |
+
},
|
| 189 |
+
"151666": {
|
| 190 |
+
"content": "</tool_response>",
|
| 191 |
+
"lstrip": false,
|
| 192 |
+
"normalized": false,
|
| 193 |
+
"rstrip": false,
|
| 194 |
+
"single_word": false,
|
| 195 |
+
"special": false
|
| 196 |
+
},
|
| 197 |
+
"151667": {
|
| 198 |
+
"content": "<think>",
|
| 199 |
+
"lstrip": false,
|
| 200 |
+
"normalized": false,
|
| 201 |
+
"rstrip": false,
|
| 202 |
+
"single_word": false,
|
| 203 |
+
"special": false
|
| 204 |
+
},
|
| 205 |
+
"151668": {
|
| 206 |
+
"content": "</think>",
|
| 207 |
+
"lstrip": false,
|
| 208 |
+
"normalized": false,
|
| 209 |
+
"rstrip": false,
|
| 210 |
+
"single_word": false,
|
| 211 |
+
"special": false
|
| 212 |
+
}
|
| 213 |
+
},
|
| 214 |
+
"additional_special_tokens": [
|
| 215 |
+
"<|im_start|>",
|
| 216 |
+
"<|im_end|>",
|
| 217 |
+
"<|object_ref_start|>",
|
| 218 |
+
"<|object_ref_end|>",
|
| 219 |
+
"<|box_start|>",
|
| 220 |
+
"<|box_end|>",
|
| 221 |
+
"<|quad_start|>",
|
| 222 |
+
"<|quad_end|>",
|
| 223 |
+
"<|vision_start|>",
|
| 224 |
+
"<|vision_end|>",
|
| 225 |
+
"<|vision_pad|>",
|
| 226 |
+
"<|image_pad|>",
|
| 227 |
+
"<|video_pad|>"
|
| 228 |
+
],
|
| 229 |
+
"bos_token": null,
|
| 230 |
+
"clean_up_tokenization_spaces": false,
|
| 231 |
+
"eos_token": "<|im_end|>",
|
| 232 |
+
"errors": "replace",
|
| 233 |
+
"extra_special_tokens": {},
|
| 234 |
+
"model_max_length": 131072,
|
| 235 |
+
"pad_token": "<|endoftext|>",
|
| 236 |
+
"split_special_tokens": false,
|
| 237 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 238 |
+
"unk_token": null
|
| 239 |
+
}
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|