GadflyII commited on
Commit
0cc906e
·
verified ·
1 Parent(s): 2063dd3

Upload folder using huggingface_hub

Browse files
.gitattributes CHANGED
@@ -33,3 +33,4 @@ 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
 
 
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
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,230 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ language:
4
+ - en
5
+ - zh
6
+ base_model: zai-org/GLM-4.7-Flash
7
+ tags:
8
+ - moe
9
+ - mxfp4
10
+ - quantized
11
+ - vllm
12
+ - glm
13
+ - 30b
14
+ library_name: transformers
15
+ pipeline_tag: text-generation
16
+ ---
17
+ # Note: If you have a multi-GPU SM120 Blackwell system (RTX 50/Pro), try my vLLM fork to resolve P2P / TP=2 issues (Pending PR into upstream).
18
+ https://github.com/Gadflyii/vllm/tree/main
19
+
20
+ # GLM-4.7-Flash MXFP4
21
+
22
+ This is a **MXFP4 quantization** of [zai-org/GLM-4.7-Flash](https://huggingface.co/zai-org/GLM-4.7-Flash), a 30B-A3B (30B total, 3B active) Mixture-of-Experts model.
23
+
24
+ ## Quantization Strategy
25
+
26
+ This model uses **MXFP4 (Microscaling FP4)** format with the Marlin backend for inference. Custom quantization with calibration (128 samples, 2048 max seq len) applied to MoE experts.
27
+
28
+ | Component | Precision | Rationale |
29
+ |-----------|-----------|-----------|
30
+ | MLP Experts (gate_up, down) | MXFP4 (E2M1) | 64 routed experts, 4 active per token |
31
+ | **Attention (MLA)** | **BF16** | Low-rank compressed Q/KV projections are sensitive |
32
+ | Dense MLP | BF16 | First layer dense MLP |
33
+ | Norms, Gates, Embeddings | BF16 | Standard practice |
34
+
35
+ ### MXFP4 vs NVFP4
36
+
37
+ | Property | MXFP4 | NVFP4 |
38
+ |----------|-------|-------|
39
+ | Weight Format | E2M1 (4-bit) | E2M1 (4-bit) |
40
+ | Scale Format | E8M0 (power-of-2) | FP8 (E4M3) |
41
+ | Block Size | 32 | 16 |
42
+ | Backend | Marlin | FlashInfer/Cutlass |
43
+
44
+ ## Performance
45
+
46
+ | Metric | BF16 | **This Model** |
47
+ |--------|------|----------------|
48
+ | MMLU-Pro | 24.83% | **25.86%** |
49
+ | Size | 62.4 GB | **20.8 GB** |
50
+ | Compression | 1x | **3.0x** |
51
+ | Accuracy Δ | - | **+1.03%** |
52
+ | Throughput | ~92 tok/s | **~140 tok/s** |
53
+ | Max Context (single 96GB GPU) | ~26K | **~69K** |
54
+
55
+ ## Usage
56
+
57
+ ### Requirements
58
+
59
+ - **vLLM**: 0.14.0+ (for MXFP4 Marlin backend support)
60
+ - **transformers**: 5.0.0+ (for `glm4_moe_lite` architecture)
61
+ - **GPU**: NVIDIA GPU with compute capability 8.0+ (Ampere/Hopper/Blackwell)
62
+
63
+ ### Installation
64
+
65
+ ```bash
66
+ pip install vllm>=0.14.0
67
+ pip install git+https://github.com/huggingface/transformers.git
68
+ ```
69
+
70
+ ### Inference with vLLM
71
+
72
+ ```python
73
+ import os
74
+ os.environ["VLLM_MXFP4_USE_MARLIN"] = "1"
75
+
76
+ from vllm import LLM, SamplingParams
77
+
78
+ model = LLM(
79
+ "GadflyII/GLM-4.7-Flash-MXFP4",
80
+ tensor_parallel_size=1,
81
+ max_model_len=65536, # Can go up to 202K with sufficient VRAM
82
+ trust_remote_code=True,
83
+ gpu_memory_utilization=0.90,
84
+ )
85
+
86
+ # Note: Do NOT use repetition_penalty > 1.05, it causes degradation at long outputs
87
+ params = SamplingParams(temperature=0.7, top_p=0.9, max_tokens=2048)
88
+ outputs = model.generate(["Explain quantum computing in simple terms."], params)
89
+ print(outputs[0].outputs[0].text)
90
+ ```
91
+
92
+ ### Serving with vLLM
93
+
94
+ ```bash
95
+ VLLM_MXFP4_USE_MARLIN=1 vllm serve GadflyII/GLM-4.7-Flash-MXFP4 \
96
+ --tensor-parallel-size 1 \
97
+ --max-model-len 65536 \
98
+ --trust-remote-code \
99
+ --gpu-memory-utilization 0.90
100
+ ```
101
+
102
+ ### Chat Completions API
103
+
104
+ ```python
105
+ import requests
106
+
107
+ payload = {
108
+ "model": "GadflyII/GLM-4.7-Flash-MXFP4",
109
+ "messages": [{"role": "user", "content": "Hello!"}],
110
+ "max_tokens": 1024,
111
+ "temperature": 0.7,
112
+ # Disable thinking mode for direct responses:
113
+ "chat_template_kwargs": {"enable_thinking": False}
114
+ # Or enable thinking for reasoning tasks:
115
+ # "chat_template_kwargs": {"enable_thinking": True}
116
+ }
117
+ response = requests.post("http://localhost:8000/v1/chat/completions", json=payload)
118
+ print(response.json()["choices"][0]["message"]["content"])
119
+ ```
120
+
121
+ ## Important Usage Notes
122
+
123
+ ### Sampling Parameters
124
+
125
+ | Parameter | Recommended | Avoid | Reason |
126
+ |-----------|-------------|-------|--------|
127
+ | `temperature` | 0.3-0.7 | - | Standard range |
128
+ | `top_p` | 0.9-0.95 | - | Standard range |
129
+ | `repetition_penalty` | None or ≤1.05 | >1.05 | High values cause word-salad at long outputs |
130
+ | `max_tokens` | Up to 10,000+ | - | Model handles long generation well |
131
+
132
+ ### Thinking Mode
133
+
134
+ This model supports a "thinking" mode where it shows its reasoning process:
135
+
136
+ - **`enable_thinking: True`** - Model outputs its reasoning process before the answer (good for math, coding, complex reasoning)
137
+ - **`enable_thinking: False`** - Model outputs the answer directly (good for chat, simple Q&A)
138
+
139
+ The model thinks in English when given English prompts.
140
+
141
+ ### Long Context
142
+
143
+ With 3x compression, this model leaves significantly more VRAM for KV cache compared to BF16:
144
+
145
+ | GPU VRAM | BF16 Max Context | MXFP4 Max Context |
146
+ |----------|------------------|-------------------|
147
+ | 24 GB | ~2K | ~8K |
148
+ | 48 GB | ~12K | ~32K |
149
+ | 80 GB | ~24K | ~60K |
150
+ | 96 GB | ~26K | ~69K |
151
+
152
+ ## Model Details
153
+
154
+ - **Base Model**: [zai-org/GLM-4.7-Flash](https://huggingface.co/zai-org/GLM-4.7-Flash)
155
+ - **Architecture**: `Glm4MoeLiteForCausalLM`
156
+ - **Parameters**: 30B total, 3B active per token (30B-A3B)
157
+ - **MoE Configuration**: 64 routed experts, 4 active, 1 shared expert
158
+ - **Layers**: 47
159
+ - **Context Length**: 202,752 tokens (max)
160
+ - **Languages**: English, Chinese
161
+
162
+ ## Quantization Details
163
+
164
+ - **Format**: MXFP4 (Microscaling FP4)
165
+ - **Weight Format**: E2M1 (4-bit floating point, range ±6.0)
166
+ - **Scale Format**: E8M0 (8-bit power-of-2 scales)
167
+ - **Block Size**: 32
168
+ - **Calibration**: 128 samples from neuralmagic/calibration dataset
169
+
170
+ ## Evaluation
171
+
172
+ ### MMLU-Pro Overall Results
173
+
174
+ | Model | Accuracy | Correct | Total |
175
+ |-------|----------|---------|-------|
176
+ | **BF16 (baseline)** | 24.83% | 2988 | 12032 |
177
+ | **MXFP4 (this model)** | 25.86% | 3112 | 12032 |
178
+ | **Difference** | +1.03% | +124 | - |
179
+
180
+ ### MMLU-Pro by Category
181
+
182
+ | Category | BF16 | MXFP4 | Difference |
183
+ |----------|------|-------|------------|
184
+ | Social Sciences | 32.70% | 34.68% | +1.98% |
185
+ | Other | 31.57% | 32.84% | +1.27% |
186
+ | Humanities | 23.78% | 23.78% | 0.00% |
187
+ | STEM | 19.94% | 20.86% | +0.92% |
188
+
189
+ ### MMLU-Pro by Subject
190
+
191
+ | Subject | BF16 | MXFP4 | Difference |
192
+ |---------|------|-------|------------|
193
+ | Biology | 50.35% | 52.16% | +1.81% |
194
+ | Psychology | 44.99% | 47.74% | +2.75% |
195
+ | Economics | 36.37% | 38.27% | +1.90% |
196
+ | Health | 35.21% | 36.31% | +1.10% |
197
+ | History | 33.60% | 32.28% | -1.32% |
198
+ | Philosophy | 31.46% | 31.86% | +0.40% |
199
+ | Other | 28.35% | 29.76% | +1.41% |
200
+ | Computer Science | 26.10% | 25.85% | -0.25% |
201
+ | Business | 16.35% | 17.62% | +1.27% |
202
+ | Law | 16.89% | 17.17% | +0.28% |
203
+ | Physics | 15.32% | 16.17% | +0.85% |
204
+ | Engineering | 16.00% | 15.58% | -0.42% |
205
+ | Math | 14.06% | 15.54% | +1.48% |
206
+ | Chemistry | 14.13% | 15.46% | +1.33% |
207
+
208
+ ### Long Generation Quality
209
+
210
+ Tested generating 10,000 tokens with thinking enabled:
211
+ - **Result**: Clean, coherent output with natural stop
212
+ - **Throughput**: 72 tok/s
213
+ - **Quality**: No degradation or repetition
214
+
215
+ ## Citation
216
+
217
+ If you use this model, please cite the original GLM-4.7-Flash:
218
+
219
+ ```bibtex
220
+ @misc{glm4flash2025,
221
+ title={GLM-4.7-Flash},
222
+ author={Zhipu AI},
223
+ year={2025},
224
+ howpublished={\url{https://huggingface.co/zai-org/GLM-4.7-Flash}}
225
+ }
226
+ ```
227
+
228
+ ## License
229
+
230
+ This model inherits the Apache 2.0 license from the base model.
chat_template.jinja ADDED
@@ -0,0 +1,86 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [gMASK]<sop>
2
+ {%- if tools -%}
3
+ <|system|>
4
+ # Tools
5
+
6
+ You may call one or more functions to assist with the user query.
7
+
8
+ You are provided with function signatures within <tools></tools> XML tags:
9
+ <tools>
10
+ {% for tool in tools %}
11
+ {{ tool | tojson(ensure_ascii=False) }}
12
+ {% endfor %}
13
+ </tools>
14
+
15
+ For each function call, output the function name and arguments within the following XML format:
16
+ <tool_call>{function-name}<arg_key>{arg-key-1}</arg_key><arg_value>{arg-value-1}</arg_value><arg_key>{arg-key-2}</arg_key><arg_value>{arg-value-2}</arg_value>...</tool_call>{%- endif -%}
17
+ {%- macro visible_text(content) -%}
18
+ {%- if content is string -%}
19
+ {{- content }}
20
+ {%- elif content is iterable and content is not mapping -%}
21
+ {%- for item in content -%}
22
+ {%- if item is mapping and item.type == 'text' -%}
23
+ {{- item.text }}
24
+ {%- elif item is string -%}
25
+ {{- item }}
26
+ {%- endif -%}
27
+ {%- endfor -%}
28
+ {%- else -%}
29
+ {{- content }}
30
+ {%- endif -%}
31
+ {%- endmacro -%}
32
+ {%- set ns = namespace(last_user_index=-1) %}
33
+ {%- for m in messages %}
34
+ {%- if m.role == 'user' %}
35
+ {% set ns.last_user_index = loop.index0 -%}
36
+ {%- endif %}
37
+ {%- endfor %}
38
+ {% for m in messages %}
39
+ {%- if m.role == 'user' -%}<|user|>{{ visible_text(m.content) }}
40
+ {%- elif m.role == 'assistant' -%}
41
+ <|assistant|>
42
+ {%- set reasoning_content = '' %}
43
+ {%- set content = visible_text(m.content) %}
44
+ {%- if m.reasoning_content is string %}
45
+ {%- set reasoning_content = m.reasoning_content %}
46
+ {%- else %}
47
+ {%- if '</think>' in content %}
48
+ {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
49
+ {%- set content = content.split('</think>')[-1].lstrip('\n') %}
50
+ {%- endif %}
51
+ {%- endif %}
52
+ {%- if ((clear_thinking is defined and not clear_thinking) or loop.index0 > ns.last_user_index) and reasoning_content -%}
53
+ {{ '<think>' + reasoning_content.strip() + '</think>'}}
54
+ {%- else -%}
55
+ {{ '</think>' }}
56
+ {%- endif -%}
57
+ {%- if content.strip() -%}
58
+ {{ content.strip() }}
59
+ {%- endif -%}
60
+ {% if m.tool_calls %}
61
+ {% for tc in m.tool_calls %}
62
+ {%- if tc.function %}
63
+ {%- set tc = tc.function %}
64
+ {%- endif %}
65
+ {{- '<tool_call>' + tc.name -}}
66
+ {% set _args = tc.arguments %}{% for k, v in _args.items() %}<arg_key>{{ k }}</arg_key><arg_value>{{ v | tojson(ensure_ascii=False) if v is not string else v }}</arg_value>{% endfor %}</tool_call>{% endfor %}
67
+ {% endif %}
68
+ {%- elif m.role == 'tool' -%}
69
+ {%- if m.content is string -%}
70
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
71
+ {{- '<|observation|>' }}
72
+ {%- endif %}
73
+ {{- '<tool_response>' }}
74
+ {{- m.content }}
75
+ {{- '</tool_response>' }}
76
+ {%- else -%}
77
+ <|observation|>{% for tr in m.content %}
78
+ <tool_response>{{ tr.output if tr.output is defined else tr }}</tool_response>{% endfor -%}
79
+ {% endif -%}
80
+ {%- elif m.role == 'system' -%}
81
+ <|system|>{{ visible_text(m.content) }}
82
+ {%- endif -%}
83
+ {%- endfor -%}
84
+ {%- if add_generation_prompt -%}
85
+ <|assistant|>{{- '</think>' if (enable_thinking is defined and not enable_thinking) else '<think>' -}}
86
+ {%- endif -%}
config.json ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Glm4MoeLiteForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "pad_token_id": 154820,
8
+ "eos_token_id": [
9
+ 154820,
10
+ 154827,
11
+ 154829
12
+ ],
13
+ "hidden_act": "silu",
14
+ "hidden_size": 2048,
15
+ "intermediate_size": 10240,
16
+ "max_position_embeddings": 202752,
17
+ "model_type": "glm4_moe_lite",
18
+ "moe_intermediate_size": 1536,
19
+ "topk_method": "noaux_tc",
20
+ "norm_topk_prob": true,
21
+ "num_attention_heads": 20,
22
+ "n_group": 1,
23
+ "topk_group": 1,
24
+ "n_routed_experts": 64,
25
+ "n_shared_experts": 1,
26
+ "routed_scaling_factor": 1.8,
27
+ "num_experts_per_tok": 4,
28
+ "first_k_dense_replace": 1,
29
+ "num_hidden_layers": 47,
30
+ "num_key_value_heads": 20,
31
+ "num_nextn_predict_layers": 1,
32
+ "partial_rotary_factor": 1.0,
33
+ "rms_norm_eps": 1e-05,
34
+ "rope_scaling": null,
35
+ "rope_theta": 1000000,
36
+ "tie_word_embeddings": false,
37
+ "dtype": "bfloat16",
38
+ "transformers_version": "5.0.0rc0",
39
+ "q_lora_rank": 768,
40
+ "kv_lora_rank": 512,
41
+ "qk_nope_head_dim": 192,
42
+ "qk_rope_head_dim": 64,
43
+ "v_head_dim": 256,
44
+ "vocab_size": 154880,
45
+ "quantization_config": {
46
+ "modules_to_not_convert": [
47
+ "model.layers.*.self_attn",
48
+ "model.layers.*.mlp.gate",
49
+ "model.embed_tokens",
50
+ "lm_head"
51
+ ],
52
+ "quant_method": "mxfp4"
53
+ }
54
+ }
generation_config.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "eos_token_id": [
4
+ 154820,
5
+ 154827,
6
+ 154829
7
+ ],
8
+ "pad_token_id": 154820,
9
+ "temperature": 1.0,
10
+ "transformers_version": "5.0.0.dev0"
11
+ }
model-00000-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cec1e41cc4ba50483dc9c314b62ef2b82f65e851cea211314ace9ebaecb2dcbc
3
+ size 5358627984
model-00001-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:38b96250162478b63280eef0ddbbb2a72c53a26b8238a1d10ff711a842fc401a
3
+ size 5251542160
model-00002-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7555b79f54d6f41e672a9770f8f9773992a5bd2b267fb1f9cc735781d0293296
3
+ size 5305030152
model-00003-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d4b34764f08877889321bfe4ca4f872b438038f711e99dde3510c5f1860f39a7
3
+ size 4865219512
model.safetensors.index.json ADDED
@@ -0,0 +1,966 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {},
3
+ "weight_map": {
4
+ "model.layers.1.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
5
+ "model.layers.1.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
6
+ "model.layers.1.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
7
+ "model.layers.1.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
8
+ "model.layers.1.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
9
+ "model.layers.1.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
10
+ "model.layers.10.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
11
+ "model.layers.10.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
12
+ "model.layers.10.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
13
+ "model.layers.10.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
14
+ "model.layers.10.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
15
+ "model.layers.10.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
16
+ "model.layers.11.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
17
+ "model.layers.11.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
18
+ "model.layers.11.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
19
+ "model.layers.11.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
20
+ "model.layers.11.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
21
+ "model.layers.11.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
22
+ "model.layers.12.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
23
+ "model.layers.12.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
24
+ "model.layers.12.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
25
+ "model.layers.12.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
26
+ "model.layers.12.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
27
+ "model.layers.12.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
28
+ "model.layers.13.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
29
+ "model.layers.13.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
30
+ "model.layers.13.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
31
+ "model.layers.13.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
32
+ "model.layers.13.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
33
+ "model.layers.13.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
34
+ "model.layers.14.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
35
+ "model.layers.14.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
36
+ "model.layers.14.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
37
+ "model.layers.14.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
38
+ "model.layers.14.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
39
+ "model.layers.14.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
40
+ "model.layers.15.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
41
+ "model.layers.15.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
42
+ "model.layers.15.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
43
+ "model.layers.15.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
44
+ "model.layers.15.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
45
+ "model.layers.15.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
46
+ "model.layers.16.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
47
+ "model.layers.16.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
48
+ "model.layers.16.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
49
+ "model.layers.16.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
50
+ "model.layers.16.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
51
+ "model.layers.16.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
52
+ "model.layers.17.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
53
+ "model.layers.17.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
54
+ "model.layers.17.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
55
+ "model.layers.17.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
56
+ "model.layers.17.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
57
+ "model.layers.17.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
58
+ "model.layers.18.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
59
+ "model.layers.18.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
60
+ "model.layers.18.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
61
+ "model.layers.18.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
62
+ "model.layers.18.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
63
+ "model.layers.18.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
64
+ "model.layers.19.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
65
+ "model.layers.19.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
66
+ "model.layers.19.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
67
+ "model.layers.19.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
68
+ "model.layers.19.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
69
+ "model.layers.19.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
70
+ "model.layers.2.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
71
+ "model.layers.2.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
72
+ "model.layers.2.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
73
+ "model.layers.2.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
74
+ "model.layers.2.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
75
+ "model.layers.2.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
76
+ "model.layers.20.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
77
+ "model.layers.20.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
78
+ "model.layers.20.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
79
+ "model.layers.20.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
80
+ "model.layers.20.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
81
+ "model.layers.20.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
82
+ "model.layers.21.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
83
+ "model.layers.21.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
84
+ "model.layers.21.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
85
+ "model.layers.21.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
86
+ "model.layers.21.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
87
+ "model.layers.21.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
88
+ "model.layers.22.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
89
+ "model.layers.22.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
90
+ "model.layers.22.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
91
+ "model.layers.22.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
92
+ "model.layers.22.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
93
+ "model.layers.22.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
94
+ "model.layers.23.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
95
+ "model.layers.23.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
96
+ "model.layers.23.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
97
+ "model.layers.23.mlp.experts.w2_weight": "model-00000-of-00004.safetensors",
98
+ "model.layers.23.mlp.experts.w2_weight_scale": "model-00000-of-00004.safetensors",
99
+ "model.layers.23.mlp.experts.w2_bias": "model-00000-of-00004.safetensors",
100
+ "model.layers.24.mlp.experts.w13_weight": "model-00000-of-00004.safetensors",
101
+ "model.layers.24.mlp.experts.w13_weight_scale": "model-00000-of-00004.safetensors",
102
+ "model.layers.24.mlp.experts.w13_bias": "model-00000-of-00004.safetensors",
103
+ "model.layers.24.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
104
+ "model.layers.24.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
105
+ "model.layers.24.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
106
+ "model.layers.25.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
107
+ "model.layers.25.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
108
+ "model.layers.25.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
109
+ "model.layers.25.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
110
+ "model.layers.25.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
111
+ "model.layers.25.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
112
+ "model.layers.26.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
113
+ "model.layers.26.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
114
+ "model.layers.26.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
115
+ "model.layers.26.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
116
+ "model.layers.26.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
117
+ "model.layers.26.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
118
+ "model.layers.27.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
119
+ "model.layers.27.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
120
+ "model.layers.27.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
121
+ "model.layers.27.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
122
+ "model.layers.27.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
123
+ "model.layers.27.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
124
+ "model.layers.28.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
125
+ "model.layers.28.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
126
+ "model.layers.28.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
127
+ "model.layers.28.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
128
+ "model.layers.28.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
129
+ "model.layers.28.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
130
+ "model.layers.29.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
131
+ "model.layers.29.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
132
+ "model.layers.29.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
133
+ "model.layers.29.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
134
+ "model.layers.29.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
135
+ "model.layers.29.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
136
+ "model.layers.3.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
137
+ "model.layers.3.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
138
+ "model.layers.3.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
139
+ "model.layers.3.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
140
+ "model.layers.3.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
141
+ "model.layers.3.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
142
+ "model.layers.30.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
143
+ "model.layers.30.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
144
+ "model.layers.30.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
145
+ "model.layers.30.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
146
+ "model.layers.30.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
147
+ "model.layers.30.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
148
+ "model.layers.31.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
149
+ "model.layers.31.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
150
+ "model.layers.31.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
151
+ "model.layers.31.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
152
+ "model.layers.31.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
153
+ "model.layers.31.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
154
+ "model.layers.32.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
155
+ "model.layers.32.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
156
+ "model.layers.32.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
157
+ "model.layers.32.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
158
+ "model.layers.32.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
159
+ "model.layers.32.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
160
+ "model.layers.33.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
161
+ "model.layers.33.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
162
+ "model.layers.33.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
163
+ "model.layers.33.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
164
+ "model.layers.33.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
165
+ "model.layers.33.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
166
+ "model.layers.34.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
167
+ "model.layers.34.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
168
+ "model.layers.34.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
169
+ "model.layers.34.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
170
+ "model.layers.34.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
171
+ "model.layers.34.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
172
+ "model.layers.35.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
173
+ "model.layers.35.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
174
+ "model.layers.35.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
175
+ "model.layers.35.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
176
+ "model.layers.35.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
177
+ "model.layers.35.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
178
+ "model.layers.36.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
179
+ "model.layers.36.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
180
+ "model.layers.36.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
181
+ "model.layers.36.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
182
+ "model.layers.36.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
183
+ "model.layers.36.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
184
+ "model.layers.37.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
185
+ "model.layers.37.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
186
+ "model.layers.37.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
187
+ "model.layers.37.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
188
+ "model.layers.37.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
189
+ "model.layers.37.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
190
+ "model.layers.38.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
191
+ "model.layers.38.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
192
+ "model.layers.38.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
193
+ "model.layers.38.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
194
+ "model.layers.38.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
195
+ "model.layers.38.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
196
+ "model.layers.39.mlp.experts.w13_weight": "model-00001-of-00004.safetensors",
197
+ "model.layers.39.mlp.experts.w13_weight_scale": "model-00001-of-00004.safetensors",
198
+ "model.layers.39.mlp.experts.w13_bias": "model-00001-of-00004.safetensors",
199
+ "model.layers.39.mlp.experts.w2_weight": "model-00001-of-00004.safetensors",
200
+ "model.layers.39.mlp.experts.w2_weight_scale": "model-00001-of-00004.safetensors",
201
+ "model.layers.39.mlp.experts.w2_bias": "model-00001-of-00004.safetensors",
202
+ "model.layers.4.mlp.experts.w13_weight": "model-00002-of-00004.safetensors",
203
+ "model.layers.4.mlp.experts.w13_weight_scale": "model-00002-of-00004.safetensors",
204
+ "model.layers.4.mlp.experts.w13_bias": "model-00002-of-00004.safetensors",
205
+ "model.layers.4.mlp.experts.w2_weight": "model-00002-of-00004.safetensors",
206
+ "model.layers.4.mlp.experts.w2_weight_scale": "model-00002-of-00004.safetensors",
207
+ "model.layers.4.mlp.experts.w2_bias": "model-00002-of-00004.safetensors",
208
+ "model.layers.40.mlp.experts.w13_weight": "model-00002-of-00004.safetensors",
209
+ "model.layers.40.mlp.experts.w13_weight_scale": "model-00002-of-00004.safetensors",
210
+ "model.layers.40.mlp.experts.w13_bias": "model-00002-of-00004.safetensors",
211
+ "model.layers.40.mlp.experts.w2_weight": "model-00002-of-00004.safetensors",
212
+ "model.layers.40.mlp.experts.w2_weight_scale": "model-00002-of-00004.safetensors",
213
+ "model.layers.40.mlp.experts.w2_bias": "model-00002-of-00004.safetensors",
214
+ "model.layers.41.mlp.experts.w13_weight": "model-00002-of-00004.safetensors",
215
+ "model.layers.41.mlp.experts.w13_weight_scale": "model-00002-of-00004.safetensors",
216
+ "model.layers.41.mlp.experts.w13_bias": "model-00002-of-00004.safetensors",
217
+ "model.layers.41.mlp.experts.w2_weight": "model-00002-of-00004.safetensors",
218
+ "model.layers.41.mlp.experts.w2_weight_scale": "model-00002-of-00004.safetensors",
219
+ "model.layers.41.mlp.experts.w2_bias": "model-00002-of-00004.safetensors",
220
+ "model.layers.42.mlp.experts.w13_weight": "model-00002-of-00004.safetensors",
221
+ "model.layers.42.mlp.experts.w13_weight_scale": "model-00002-of-00004.safetensors",
222
+ "model.layers.42.mlp.experts.w13_bias": "model-00002-of-00004.safetensors",
223
+ "model.layers.42.mlp.experts.w2_weight": "model-00002-of-00004.safetensors",
224
+ "model.layers.42.mlp.experts.w2_weight_scale": "model-00002-of-00004.safetensors",
225
+ "model.layers.42.mlp.experts.w2_bias": "model-00002-of-00004.safetensors",
226
+ "model.layers.43.mlp.experts.w13_weight": "model-00002-of-00004.safetensors",
227
+ "model.layers.43.mlp.experts.w13_weight_scale": "model-00002-of-00004.safetensors",
228
+ "model.layers.43.mlp.experts.w13_bias": "model-00002-of-00004.safetensors",
229
+ "model.layers.43.mlp.experts.w2_weight": "model-00002-of-00004.safetensors",
230
+ "model.layers.43.mlp.experts.w2_weight_scale": "model-00002-of-00004.safetensors",
231
+ "model.layers.43.mlp.experts.w2_bias": "model-00002-of-00004.safetensors",
232
+ "model.layers.44.mlp.experts.w13_weight": "model-00002-of-00004.safetensors",
233
+ "model.layers.44.mlp.experts.w13_weight_scale": "model-00002-of-00004.safetensors",
234
+ "model.layers.44.mlp.experts.w13_bias": "model-00002-of-00004.safetensors",
235
+ "model.layers.44.mlp.experts.w2_weight": "model-00002-of-00004.safetensors",
236
+ "model.layers.44.mlp.experts.w2_weight_scale": "model-00002-of-00004.safetensors",
237
+ "model.layers.44.mlp.experts.w2_bias": "model-00002-of-00004.safetensors",
238
+ "model.layers.45.mlp.experts.w13_weight": "model-00002-of-00004.safetensors",
239
+ "model.layers.45.mlp.experts.w13_weight_scale": "model-00002-of-00004.safetensors",
240
+ "model.layers.45.mlp.experts.w13_bias": "model-00002-of-00004.safetensors",
241
+ "model.layers.45.mlp.experts.w2_weight": "model-00002-of-00004.safetensors",
242
+ "model.layers.45.mlp.experts.w2_weight_scale": "model-00002-of-00004.safetensors",
243
+ "model.layers.45.mlp.experts.w2_bias": "model-00002-of-00004.safetensors",
244
+ "model.layers.46.mlp.experts.w13_weight": "model-00002-of-00004.safetensors",
245
+ "model.layers.46.mlp.experts.w13_weight_scale": "model-00002-of-00004.safetensors",
246
+ "model.layers.46.mlp.experts.w13_bias": "model-00002-of-00004.safetensors",
247
+ "model.layers.46.mlp.experts.w2_weight": "model-00002-of-00004.safetensors",
248
+ "model.layers.46.mlp.experts.w2_weight_scale": "model-00002-of-00004.safetensors",
249
+ "model.layers.46.mlp.experts.w2_bias": "model-00002-of-00004.safetensors",
250
+ "model.layers.47.mlp.experts.w13_weight": "model-00002-of-00004.safetensors",
251
+ "model.layers.47.mlp.experts.w13_weight_scale": "model-00002-of-00004.safetensors",
252
+ "model.layers.47.mlp.experts.w13_bias": "model-00002-of-00004.safetensors",
253
+ "model.layers.47.mlp.experts.w2_weight": "model-00002-of-00004.safetensors",
254
+ "model.layers.47.mlp.experts.w2_weight_scale": "model-00002-of-00004.safetensors",
255
+ "model.layers.47.mlp.experts.w2_bias": "model-00002-of-00004.safetensors",
256
+ "model.layers.5.mlp.experts.w13_weight": "model-00002-of-00004.safetensors",
257
+ "model.layers.5.mlp.experts.w13_weight_scale": "model-00002-of-00004.safetensors",
258
+ "model.layers.5.mlp.experts.w13_bias": "model-00002-of-00004.safetensors",
259
+ "model.layers.5.mlp.experts.w2_weight": "model-00002-of-00004.safetensors",
260
+ "model.layers.5.mlp.experts.w2_weight_scale": "model-00002-of-00004.safetensors",
261
+ "model.layers.5.mlp.experts.w2_bias": "model-00002-of-00004.safetensors",
262
+ "model.layers.6.mlp.experts.w13_weight": "model-00002-of-00004.safetensors",
263
+ "model.layers.6.mlp.experts.w13_weight_scale": "model-00002-of-00004.safetensors",
264
+ "model.layers.6.mlp.experts.w13_bias": "model-00002-of-00004.safetensors",
265
+ "model.layers.6.mlp.experts.w2_weight": "model-00002-of-00004.safetensors",
266
+ "model.layers.6.mlp.experts.w2_weight_scale": "model-00002-of-00004.safetensors",
267
+ "model.layers.6.mlp.experts.w2_bias": "model-00002-of-00004.safetensors",
268
+ "model.layers.7.mlp.experts.w13_weight": "model-00002-of-00004.safetensors",
269
+ "model.layers.7.mlp.experts.w13_weight_scale": "model-00002-of-00004.safetensors",
270
+ "model.layers.7.mlp.experts.w13_bias": "model-00002-of-00004.safetensors",
271
+ "model.layers.7.mlp.experts.w2_weight": "model-00002-of-00004.safetensors",
272
+ "model.layers.7.mlp.experts.w2_weight_scale": "model-00002-of-00004.safetensors",
273
+ "model.layers.7.mlp.experts.w2_bias": "model-00002-of-00004.safetensors",
274
+ "model.layers.8.mlp.experts.w13_weight": "model-00002-of-00004.safetensors",
275
+ "model.layers.8.mlp.experts.w13_weight_scale": "model-00002-of-00004.safetensors",
276
+ "model.layers.8.mlp.experts.w13_bias": "model-00002-of-00004.safetensors",
277
+ "model.layers.8.mlp.experts.w2_weight": "model-00002-of-00004.safetensors",
278
+ "model.layers.8.mlp.experts.w2_weight_scale": "model-00002-of-00004.safetensors",
279
+ "model.layers.8.mlp.experts.w2_bias": "model-00002-of-00004.safetensors",
280
+ "model.layers.9.mlp.experts.w13_weight": "model-00002-of-00004.safetensors",
281
+ "model.layers.9.mlp.experts.w13_weight_scale": "model-00002-of-00004.safetensors",
282
+ "model.layers.9.mlp.experts.w13_bias": "model-00002-of-00004.safetensors",
283
+ "model.layers.9.mlp.experts.w2_weight": "model-00002-of-00004.safetensors",
284
+ "model.layers.9.mlp.experts.w2_weight_scale": "model-00002-of-00004.safetensors",
285
+ "model.layers.9.mlp.experts.w2_bias": "model-00002-of-00004.safetensors",
286
+ "model.embed_tokens.weight": "model-00002-of-00004.safetensors",
287
+ "model.layers.0.input_layernorm.weight": "model-00002-of-00004.safetensors",
288
+ "model.layers.0.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
289
+ "model.layers.0.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
290
+ "model.layers.0.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
291
+ "model.layers.0.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
292
+ "model.layers.0.self_attn.kv_a_layernorm.weight": "model-00002-of-00004.safetensors",
293
+ "model.layers.0.self_attn.kv_a_proj_with_mqa.weight": "model-00002-of-00004.safetensors",
294
+ "model.layers.0.self_attn.kv_b_proj.weight": "model-00002-of-00004.safetensors",
295
+ "model.layers.0.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
296
+ "model.layers.0.self_attn.q_a_layernorm.weight": "model-00002-of-00004.safetensors",
297
+ "model.layers.0.self_attn.q_a_proj.weight": "model-00002-of-00004.safetensors",
298
+ "model.layers.0.self_attn.q_b_proj.weight": "model-00002-of-00004.safetensors",
299
+ "model.layers.47.embed_tokens.weight": "model-00003-of-00004.safetensors",
300
+ "model.layers.1.input_layernorm.weight": "model-00003-of-00004.safetensors",
301
+ "model.layers.1.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
302
+ "model.layers.1.mlp.gate.weight": "model-00003-of-00004.safetensors",
303
+ "model.layers.1.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
304
+ "model.layers.1.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
305
+ "model.layers.1.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
306
+ "model.layers.1.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
307
+ "model.layers.1.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
308
+ "model.layers.1.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
309
+ "model.layers.1.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
310
+ "model.layers.1.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
311
+ "model.layers.1.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
312
+ "model.layers.1.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
313
+ "model.layers.1.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
314
+ "model.layers.2.input_layernorm.weight": "model-00003-of-00004.safetensors",
315
+ "model.layers.2.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
316
+ "model.layers.2.mlp.gate.weight": "model-00003-of-00004.safetensors",
317
+ "model.layers.2.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
318
+ "model.layers.2.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
319
+ "model.layers.2.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
320
+ "model.layers.2.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
321
+ "model.layers.2.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
322
+ "model.layers.2.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
323
+ "model.layers.2.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
324
+ "model.layers.2.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
325
+ "model.layers.2.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
326
+ "model.layers.2.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
327
+ "model.layers.2.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
328
+ "model.layers.3.input_layernorm.weight": "model-00003-of-00004.safetensors",
329
+ "model.layers.3.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
330
+ "model.layers.3.mlp.gate.weight": "model-00003-of-00004.safetensors",
331
+ "model.layers.3.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
332
+ "model.layers.3.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
333
+ "model.layers.3.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
334
+ "model.layers.3.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
335
+ "model.layers.3.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
336
+ "model.layers.3.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
337
+ "model.layers.3.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
338
+ "model.layers.3.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
339
+ "model.layers.3.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
340
+ "model.layers.3.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
341
+ "model.layers.3.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
342
+ "model.layers.4.input_layernorm.weight": "model-00003-of-00004.safetensors",
343
+ "model.layers.4.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
344
+ "model.layers.4.mlp.gate.weight": "model-00003-of-00004.safetensors",
345
+ "model.layers.4.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
346
+ "model.layers.4.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
347
+ "model.layers.4.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
348
+ "model.layers.4.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
349
+ "model.layers.4.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
350
+ "model.layers.4.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
351
+ "model.layers.4.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
352
+ "model.layers.4.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
353
+ "model.layers.4.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
354
+ "model.layers.4.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
355
+ "model.layers.4.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
356
+ "model.layers.5.input_layernorm.weight": "model-00003-of-00004.safetensors",
357
+ "model.layers.5.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
358
+ "model.layers.5.mlp.gate.weight": "model-00003-of-00004.safetensors",
359
+ "model.layers.5.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
360
+ "model.layers.5.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
361
+ "model.layers.5.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
362
+ "model.layers.5.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
363
+ "model.layers.5.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
364
+ "model.layers.5.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
365
+ "model.layers.5.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
366
+ "model.layers.5.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
367
+ "model.layers.5.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
368
+ "model.layers.5.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
369
+ "model.layers.5.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
370
+ "model.layers.6.input_layernorm.weight": "model-00003-of-00004.safetensors",
371
+ "model.layers.6.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
372
+ "model.layers.6.mlp.gate.weight": "model-00003-of-00004.safetensors",
373
+ "model.layers.6.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
374
+ "model.layers.6.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
375
+ "model.layers.6.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
376
+ "model.layers.6.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
377
+ "model.layers.6.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
378
+ "model.layers.6.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
379
+ "model.layers.6.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
380
+ "model.layers.6.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
381
+ "model.layers.6.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
382
+ "model.layers.6.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
383
+ "model.layers.6.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
384
+ "model.layers.7.input_layernorm.weight": "model-00003-of-00004.safetensors",
385
+ "model.layers.7.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
386
+ "model.layers.7.mlp.gate.weight": "model-00003-of-00004.safetensors",
387
+ "model.layers.7.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
388
+ "model.layers.7.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
389
+ "model.layers.7.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
390
+ "model.layers.7.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
391
+ "model.layers.7.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
392
+ "model.layers.7.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
393
+ "model.layers.7.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
394
+ "model.layers.7.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
395
+ "model.layers.7.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
396
+ "model.layers.7.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
397
+ "model.layers.7.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
398
+ "model.layers.8.input_layernorm.weight": "model-00003-of-00004.safetensors",
399
+ "model.layers.8.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
400
+ "model.layers.8.mlp.gate.weight": "model-00003-of-00004.safetensors",
401
+ "model.layers.8.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
402
+ "model.layers.8.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
403
+ "model.layers.8.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
404
+ "model.layers.8.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
405
+ "model.layers.8.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
406
+ "model.layers.8.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
407
+ "model.layers.8.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
408
+ "model.layers.8.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
409
+ "model.layers.8.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
410
+ "model.layers.8.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
411
+ "model.layers.8.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
412
+ "model.layers.9.input_layernorm.weight": "model-00003-of-00004.safetensors",
413
+ "model.layers.9.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
414
+ "model.layers.9.mlp.gate.weight": "model-00003-of-00004.safetensors",
415
+ "model.layers.9.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
416
+ "model.layers.9.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
417
+ "model.layers.9.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
418
+ "model.layers.9.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
419
+ "model.layers.9.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
420
+ "model.layers.9.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
421
+ "model.layers.9.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
422
+ "model.layers.9.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
423
+ "model.layers.9.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
424
+ "model.layers.9.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
425
+ "model.layers.9.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
426
+ "model.layers.10.input_layernorm.weight": "model-00003-of-00004.safetensors",
427
+ "model.layers.10.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
428
+ "model.layers.10.mlp.gate.weight": "model-00003-of-00004.safetensors",
429
+ "model.layers.10.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
430
+ "model.layers.10.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
431
+ "model.layers.10.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
432
+ "model.layers.10.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
433
+ "model.layers.10.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
434
+ "model.layers.10.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
435
+ "model.layers.10.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
436
+ "model.layers.10.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
437
+ "model.layers.10.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
438
+ "model.layers.10.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
439
+ "model.layers.10.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
440
+ "model.layers.11.input_layernorm.weight": "model-00003-of-00004.safetensors",
441
+ "model.layers.11.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
442
+ "model.layers.11.mlp.gate.weight": "model-00003-of-00004.safetensors",
443
+ "model.layers.11.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
444
+ "model.layers.11.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
445
+ "model.layers.11.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
446
+ "model.layers.11.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
447
+ "model.layers.11.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
448
+ "model.layers.11.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
449
+ "model.layers.11.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
450
+ "model.layers.11.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
451
+ "model.layers.11.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
452
+ "model.layers.11.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
453
+ "model.layers.11.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
454
+ "model.layers.12.input_layernorm.weight": "model-00003-of-00004.safetensors",
455
+ "model.layers.12.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
456
+ "model.layers.12.mlp.gate.weight": "model-00003-of-00004.safetensors",
457
+ "model.layers.12.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
458
+ "model.layers.12.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
459
+ "model.layers.12.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
460
+ "model.layers.12.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
461
+ "model.layers.12.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
462
+ "model.layers.12.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
463
+ "model.layers.12.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
464
+ "model.layers.12.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
465
+ "model.layers.12.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
466
+ "model.layers.12.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
467
+ "model.layers.12.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
468
+ "model.layers.13.input_layernorm.weight": "model-00003-of-00004.safetensors",
469
+ "model.layers.13.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
470
+ "model.layers.13.mlp.gate.weight": "model-00003-of-00004.safetensors",
471
+ "model.layers.13.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
472
+ "model.layers.13.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
473
+ "model.layers.13.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
474
+ "model.layers.13.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
475
+ "model.layers.13.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
476
+ "model.layers.13.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
477
+ "model.layers.13.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
478
+ "model.layers.13.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
479
+ "model.layers.13.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
480
+ "model.layers.13.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
481
+ "model.layers.13.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
482
+ "model.layers.14.input_layernorm.weight": "model-00003-of-00004.safetensors",
483
+ "model.layers.14.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
484
+ "model.layers.14.mlp.gate.weight": "model-00003-of-00004.safetensors",
485
+ "model.layers.14.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
486
+ "model.layers.14.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
487
+ "model.layers.14.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
488
+ "model.layers.14.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
489
+ "model.layers.14.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
490
+ "model.layers.14.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
491
+ "model.layers.14.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
492
+ "model.layers.14.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
493
+ "model.layers.14.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
494
+ "model.layers.14.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
495
+ "model.layers.14.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
496
+ "model.layers.15.input_layernorm.weight": "model-00003-of-00004.safetensors",
497
+ "model.layers.15.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
498
+ "model.layers.15.mlp.gate.weight": "model-00003-of-00004.safetensors",
499
+ "model.layers.15.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
500
+ "model.layers.15.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
501
+ "model.layers.15.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
502
+ "model.layers.15.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
503
+ "model.layers.15.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
504
+ "model.layers.15.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
505
+ "model.layers.15.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
506
+ "model.layers.15.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
507
+ "model.layers.15.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
508
+ "model.layers.15.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
509
+ "model.layers.15.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
510
+ "model.layers.16.input_layernorm.weight": "model-00003-of-00004.safetensors",
511
+ "model.layers.16.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
512
+ "model.layers.16.mlp.gate.weight": "model-00003-of-00004.safetensors",
513
+ "model.layers.16.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
514
+ "model.layers.16.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
515
+ "model.layers.16.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
516
+ "model.layers.16.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
517
+ "model.layers.16.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
518
+ "model.layers.16.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
519
+ "model.layers.16.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
520
+ "model.layers.16.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
521
+ "model.layers.16.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
522
+ "model.layers.16.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
523
+ "model.layers.16.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
524
+ "model.layers.17.input_layernorm.weight": "model-00003-of-00004.safetensors",
525
+ "model.layers.17.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
526
+ "model.layers.17.mlp.gate.weight": "model-00003-of-00004.safetensors",
527
+ "model.layers.17.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
528
+ "model.layers.17.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
529
+ "model.layers.17.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
530
+ "model.layers.17.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
531
+ "model.layers.17.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
532
+ "model.layers.17.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
533
+ "model.layers.17.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
534
+ "model.layers.17.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
535
+ "model.layers.17.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
536
+ "model.layers.17.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
537
+ "model.layers.17.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
538
+ "model.layers.18.input_layernorm.weight": "model-00003-of-00004.safetensors",
539
+ "model.layers.18.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
540
+ "model.layers.18.mlp.gate.weight": "model-00003-of-00004.safetensors",
541
+ "model.layers.18.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
542
+ "model.layers.18.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
543
+ "model.layers.18.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
544
+ "model.layers.18.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
545
+ "model.layers.18.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
546
+ "model.layers.18.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
547
+ "model.layers.18.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
548
+ "model.layers.18.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
549
+ "model.layers.18.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
550
+ "model.layers.18.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
551
+ "model.layers.18.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
552
+ "model.layers.19.input_layernorm.weight": "model-00003-of-00004.safetensors",
553
+ "model.layers.19.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
554
+ "model.layers.19.mlp.gate.weight": "model-00003-of-00004.safetensors",
555
+ "model.layers.19.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
556
+ "model.layers.19.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
557
+ "model.layers.19.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
558
+ "model.layers.19.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
559
+ "model.layers.19.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
560
+ "model.layers.19.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
561
+ "model.layers.19.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
562
+ "model.layers.19.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
563
+ "model.layers.19.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
564
+ "model.layers.19.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
565
+ "model.layers.19.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
566
+ "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors",
567
+ "model.layers.20.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
568
+ "model.layers.20.mlp.gate.weight": "model-00003-of-00004.safetensors",
569
+ "model.layers.20.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
570
+ "model.layers.20.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
571
+ "model.layers.20.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
572
+ "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
573
+ "model.layers.20.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
574
+ "model.layers.20.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
575
+ "model.layers.20.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
576
+ "model.layers.20.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
577
+ "model.layers.20.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
578
+ "model.layers.20.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
579
+ "model.layers.20.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
580
+ "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors",
581
+ "model.layers.21.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
582
+ "model.layers.21.mlp.gate.weight": "model-00003-of-00004.safetensors",
583
+ "model.layers.21.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
584
+ "model.layers.21.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
585
+ "model.layers.21.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
586
+ "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
587
+ "model.layers.21.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
588
+ "model.layers.21.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
589
+ "model.layers.21.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
590
+ "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
591
+ "model.layers.21.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
592
+ "model.layers.21.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
593
+ "model.layers.21.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
594
+ "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors",
595
+ "model.layers.22.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
596
+ "model.layers.22.mlp.gate.weight": "model-00003-of-00004.safetensors",
597
+ "model.layers.22.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
598
+ "model.layers.22.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
599
+ "model.layers.22.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
600
+ "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
601
+ "model.layers.22.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
602
+ "model.layers.22.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
603
+ "model.layers.22.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
604
+ "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
605
+ "model.layers.22.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
606
+ "model.layers.22.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
607
+ "model.layers.22.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
608
+ "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors",
609
+ "model.layers.23.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
610
+ "model.layers.23.mlp.gate.weight": "model-00003-of-00004.safetensors",
611
+ "model.layers.23.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
612
+ "model.layers.23.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
613
+ "model.layers.23.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
614
+ "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
615
+ "model.layers.23.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
616
+ "model.layers.23.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
617
+ "model.layers.23.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
618
+ "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
619
+ "model.layers.23.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
620
+ "model.layers.23.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
621
+ "model.layers.23.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
622
+ "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors",
623
+ "model.layers.24.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
624
+ "model.layers.24.mlp.gate.weight": "model-00003-of-00004.safetensors",
625
+ "model.layers.24.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
626
+ "model.layers.24.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
627
+ "model.layers.24.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
628
+ "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
629
+ "model.layers.24.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
630
+ "model.layers.24.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
631
+ "model.layers.24.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
632
+ "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
633
+ "model.layers.24.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
634
+ "model.layers.24.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
635
+ "model.layers.24.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
636
+ "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors",
637
+ "model.layers.25.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
638
+ "model.layers.25.mlp.gate.weight": "model-00003-of-00004.safetensors",
639
+ "model.layers.25.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
640
+ "model.layers.25.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
641
+ "model.layers.25.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
642
+ "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
643
+ "model.layers.25.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
644
+ "model.layers.25.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
645
+ "model.layers.25.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
646
+ "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
647
+ "model.layers.25.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
648
+ "model.layers.25.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
649
+ "model.layers.25.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
650
+ "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors",
651
+ "model.layers.26.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
652
+ "model.layers.26.mlp.gate.weight": "model-00003-of-00004.safetensors",
653
+ "model.layers.26.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
654
+ "model.layers.26.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
655
+ "model.layers.26.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
656
+ "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
657
+ "model.layers.26.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
658
+ "model.layers.26.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
659
+ "model.layers.26.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
660
+ "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
661
+ "model.layers.26.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
662
+ "model.layers.26.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
663
+ "model.layers.26.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
664
+ "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors",
665
+ "model.layers.27.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
666
+ "model.layers.27.mlp.gate.weight": "model-00003-of-00004.safetensors",
667
+ "model.layers.27.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
668
+ "model.layers.27.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
669
+ "model.layers.27.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
670
+ "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
671
+ "model.layers.27.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
672
+ "model.layers.27.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
673
+ "model.layers.27.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
674
+ "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
675
+ "model.layers.27.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
676
+ "model.layers.27.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
677
+ "model.layers.27.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
678
+ "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors",
679
+ "model.layers.28.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
680
+ "model.layers.28.mlp.gate.weight": "model-00003-of-00004.safetensors",
681
+ "model.layers.28.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
682
+ "model.layers.28.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
683
+ "model.layers.28.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
684
+ "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
685
+ "model.layers.28.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
686
+ "model.layers.28.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
687
+ "model.layers.28.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
688
+ "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
689
+ "model.layers.28.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
690
+ "model.layers.28.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
691
+ "model.layers.28.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
692
+ "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors",
693
+ "model.layers.29.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
694
+ "model.layers.29.mlp.gate.weight": "model-00003-of-00004.safetensors",
695
+ "model.layers.29.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
696
+ "model.layers.29.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
697
+ "model.layers.29.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
698
+ "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
699
+ "model.layers.29.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
700
+ "model.layers.29.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
701
+ "model.layers.29.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
702
+ "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
703
+ "model.layers.29.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
704
+ "model.layers.29.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
705
+ "model.layers.29.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
706
+ "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors",
707
+ "model.layers.30.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
708
+ "model.layers.30.mlp.gate.weight": "model-00003-of-00004.safetensors",
709
+ "model.layers.30.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
710
+ "model.layers.30.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
711
+ "model.layers.30.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
712
+ "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
713
+ "model.layers.30.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
714
+ "model.layers.30.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
715
+ "model.layers.30.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
716
+ "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
717
+ "model.layers.30.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
718
+ "model.layers.30.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
719
+ "model.layers.30.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
720
+ "model.layers.31.input_layernorm.weight": "model-00003-of-00004.safetensors",
721
+ "model.layers.31.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
722
+ "model.layers.31.mlp.gate.weight": "model-00003-of-00004.safetensors",
723
+ "model.layers.31.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
724
+ "model.layers.31.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
725
+ "model.layers.31.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
726
+ "model.layers.31.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
727
+ "model.layers.31.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
728
+ "model.layers.31.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
729
+ "model.layers.31.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
730
+ "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
731
+ "model.layers.31.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
732
+ "model.layers.31.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
733
+ "model.layers.31.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
734
+ "model.layers.32.input_layernorm.weight": "model-00003-of-00004.safetensors",
735
+ "model.layers.32.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
736
+ "model.layers.32.mlp.gate.weight": "model-00003-of-00004.safetensors",
737
+ "model.layers.32.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
738
+ "model.layers.32.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
739
+ "model.layers.32.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
740
+ "model.layers.32.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
741
+ "model.layers.32.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
742
+ "model.layers.32.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
743
+ "model.layers.32.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
744
+ "model.layers.32.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
745
+ "model.layers.32.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
746
+ "model.layers.32.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
747
+ "model.layers.32.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
748
+ "model.layers.33.input_layernorm.weight": "model-00003-of-00004.safetensors",
749
+ "model.layers.33.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
750
+ "model.layers.33.mlp.gate.weight": "model-00003-of-00004.safetensors",
751
+ "model.layers.33.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
752
+ "model.layers.33.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
753
+ "model.layers.33.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
754
+ "model.layers.33.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
755
+ "model.layers.33.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
756
+ "model.layers.33.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
757
+ "model.layers.33.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
758
+ "model.layers.33.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
759
+ "model.layers.33.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
760
+ "model.layers.33.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
761
+ "model.layers.33.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
762
+ "model.layers.34.input_layernorm.weight": "model-00003-of-00004.safetensors",
763
+ "model.layers.34.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
764
+ "model.layers.34.mlp.gate.weight": "model-00003-of-00004.safetensors",
765
+ "model.layers.34.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
766
+ "model.layers.34.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
767
+ "model.layers.34.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
768
+ "model.layers.34.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
769
+ "model.layers.34.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
770
+ "model.layers.34.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
771
+ "model.layers.34.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
772
+ "model.layers.34.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
773
+ "model.layers.34.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
774
+ "model.layers.34.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
775
+ "model.layers.34.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
776
+ "model.layers.35.input_layernorm.weight": "model-00003-of-00004.safetensors",
777
+ "model.layers.35.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
778
+ "model.layers.35.mlp.gate.weight": "model-00003-of-00004.safetensors",
779
+ "model.layers.35.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
780
+ "model.layers.35.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
781
+ "model.layers.35.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
782
+ "model.layers.35.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
783
+ "model.layers.35.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
784
+ "model.layers.35.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
785
+ "model.layers.35.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
786
+ "model.layers.35.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
787
+ "model.layers.35.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
788
+ "model.layers.35.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
789
+ "model.layers.35.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
790
+ "model.layers.36.input_layernorm.weight": "model-00003-of-00004.safetensors",
791
+ "model.layers.36.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
792
+ "model.layers.36.mlp.gate.weight": "model-00003-of-00004.safetensors",
793
+ "model.layers.36.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
794
+ "model.layers.36.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
795
+ "model.layers.36.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
796
+ "model.layers.36.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
797
+ "model.layers.36.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
798
+ "model.layers.36.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
799
+ "model.layers.36.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
800
+ "model.layers.36.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
801
+ "model.layers.36.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
802
+ "model.layers.36.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
803
+ "model.layers.36.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
804
+ "model.layers.37.input_layernorm.weight": "model-00003-of-00004.safetensors",
805
+ "model.layers.37.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
806
+ "model.layers.37.mlp.gate.weight": "model-00003-of-00004.safetensors",
807
+ "model.layers.37.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
808
+ "model.layers.37.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
809
+ "model.layers.37.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
810
+ "model.layers.37.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
811
+ "model.layers.37.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
812
+ "model.layers.37.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
813
+ "model.layers.37.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
814
+ "model.layers.37.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
815
+ "model.layers.37.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
816
+ "model.layers.37.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
817
+ "model.layers.37.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
818
+ "model.layers.38.input_layernorm.weight": "model-00003-of-00004.safetensors",
819
+ "model.layers.38.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
820
+ "model.layers.38.mlp.gate.weight": "model-00003-of-00004.safetensors",
821
+ "model.layers.38.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
822
+ "model.layers.38.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
823
+ "model.layers.38.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
824
+ "model.layers.38.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
825
+ "model.layers.38.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
826
+ "model.layers.38.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
827
+ "model.layers.38.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
828
+ "model.layers.38.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
829
+ "model.layers.38.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
830
+ "model.layers.38.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
831
+ "model.layers.38.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
832
+ "model.layers.39.input_layernorm.weight": "model-00003-of-00004.safetensors",
833
+ "model.layers.39.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
834
+ "model.layers.39.mlp.gate.weight": "model-00003-of-00004.safetensors",
835
+ "model.layers.39.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
836
+ "model.layers.39.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
837
+ "model.layers.39.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
838
+ "model.layers.39.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
839
+ "model.layers.39.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
840
+ "model.layers.39.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
841
+ "model.layers.39.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
842
+ "model.layers.39.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
843
+ "model.layers.39.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
844
+ "model.layers.39.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
845
+ "model.layers.39.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
846
+ "model.layers.40.input_layernorm.weight": "model-00003-of-00004.safetensors",
847
+ "model.layers.40.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
848
+ "model.layers.40.mlp.gate.weight": "model-00003-of-00004.safetensors",
849
+ "model.layers.40.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
850
+ "model.layers.40.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
851
+ "model.layers.40.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
852
+ "model.layers.40.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
853
+ "model.layers.40.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
854
+ "model.layers.40.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
855
+ "model.layers.40.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
856
+ "model.layers.40.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
857
+ "model.layers.40.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
858
+ "model.layers.40.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
859
+ "model.layers.40.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
860
+ "model.layers.41.input_layernorm.weight": "model-00003-of-00004.safetensors",
861
+ "model.layers.41.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
862
+ "model.layers.41.mlp.gate.weight": "model-00003-of-00004.safetensors",
863
+ "model.layers.41.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
864
+ "model.layers.41.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
865
+ "model.layers.41.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
866
+ "model.layers.41.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
867
+ "model.layers.41.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
868
+ "model.layers.41.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
869
+ "model.layers.41.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
870
+ "model.layers.41.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
871
+ "model.layers.41.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
872
+ "model.layers.41.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
873
+ "model.layers.41.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
874
+ "model.layers.42.input_layernorm.weight": "model-00003-of-00004.safetensors",
875
+ "model.layers.42.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
876
+ "model.layers.42.mlp.gate.weight": "model-00003-of-00004.safetensors",
877
+ "model.layers.42.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
878
+ "model.layers.42.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
879
+ "model.layers.42.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
880
+ "model.layers.42.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
881
+ "model.layers.42.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
882
+ "model.layers.42.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
883
+ "model.layers.42.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
884
+ "model.layers.42.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
885
+ "model.layers.42.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
886
+ "model.layers.42.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
887
+ "model.layers.42.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
888
+ "model.layers.43.input_layernorm.weight": "model-00003-of-00004.safetensors",
889
+ "model.layers.43.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
890
+ "model.layers.43.mlp.gate.weight": "model-00003-of-00004.safetensors",
891
+ "model.layers.43.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
892
+ "model.layers.43.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
893
+ "model.layers.43.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
894
+ "model.layers.43.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
895
+ "model.layers.43.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
896
+ "model.layers.43.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
897
+ "model.layers.43.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
898
+ "model.layers.43.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
899
+ "model.layers.43.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
900
+ "model.layers.43.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
901
+ "model.layers.43.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
902
+ "model.layers.44.input_layernorm.weight": "model-00003-of-00004.safetensors",
903
+ "model.layers.44.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
904
+ "model.layers.44.mlp.gate.weight": "model-00003-of-00004.safetensors",
905
+ "model.layers.44.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
906
+ "model.layers.44.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
907
+ "model.layers.44.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
908
+ "model.layers.44.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
909
+ "model.layers.44.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
910
+ "model.layers.44.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
911
+ "model.layers.44.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
912
+ "model.layers.44.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
913
+ "model.layers.44.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
914
+ "model.layers.44.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
915
+ "model.layers.44.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
916
+ "model.layers.45.input_layernorm.weight": "model-00003-of-00004.safetensors",
917
+ "model.layers.45.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
918
+ "model.layers.45.mlp.gate.weight": "model-00003-of-00004.safetensors",
919
+ "model.layers.45.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
920
+ "model.layers.45.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
921
+ "model.layers.45.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
922
+ "model.layers.45.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
923
+ "model.layers.45.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
924
+ "model.layers.45.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
925
+ "model.layers.45.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
926
+ "model.layers.45.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
927
+ "model.layers.45.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
928
+ "model.layers.45.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
929
+ "model.layers.45.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
930
+ "lm_head.weight": "model-00003-of-00004.safetensors",
931
+ "model.layers.46.input_layernorm.weight": "model-00003-of-00004.safetensors",
932
+ "model.layers.46.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
933
+ "model.layers.46.mlp.gate.weight": "model-00003-of-00004.safetensors",
934
+ "model.layers.46.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
935
+ "model.layers.46.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
936
+ "model.layers.46.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
937
+ "model.layers.46.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
938
+ "model.layers.46.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
939
+ "model.layers.46.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
940
+ "model.layers.46.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
941
+ "model.layers.46.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
942
+ "model.layers.46.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
943
+ "model.layers.46.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
944
+ "model.layers.46.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
945
+ "model.layers.47.shared_head.head.weight": "model-00003-of-00004.safetensors",
946
+ "model.norm.weight": "model-00003-of-00004.safetensors",
947
+ "model.layers.47.eh_proj.weight": "model-00003-of-00004.safetensors",
948
+ "model.layers.47.enorm.weight": "model-00003-of-00004.safetensors",
949
+ "model.layers.47.hnorm.weight": "model-00003-of-00004.safetensors",
950
+ "model.layers.47.input_layernorm.weight": "model-00003-of-00004.safetensors",
951
+ "model.layers.47.mlp.gate.e_score_correction_bias": "model-00003-of-00004.safetensors",
952
+ "model.layers.47.mlp.gate.weight": "model-00003-of-00004.safetensors",
953
+ "model.layers.47.mlp.shared_experts.down_proj.weight": "model-00003-of-00004.safetensors",
954
+ "model.layers.47.mlp.shared_experts.gate_proj.weight": "model-00003-of-00004.safetensors",
955
+ "model.layers.47.mlp.shared_experts.up_proj.weight": "model-00003-of-00004.safetensors",
956
+ "model.layers.47.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
957
+ "model.layers.47.self_attn.kv_a_layernorm.weight": "model-00003-of-00004.safetensors",
958
+ "model.layers.47.self_attn.kv_a_proj_with_mqa.weight": "model-00003-of-00004.safetensors",
959
+ "model.layers.47.self_attn.kv_b_proj.weight": "model-00003-of-00004.safetensors",
960
+ "model.layers.47.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
961
+ "model.layers.47.self_attn.q_a_layernorm.weight": "model-00003-of-00004.safetensors",
962
+ "model.layers.47.self_attn.q_a_proj.weight": "model-00003-of-00004.safetensors",
963
+ "model.layers.47.self_attn.q_b_proj.weight": "model-00003-of-00004.safetensors",
964
+ "model.layers.47.shared_head.norm.weight": "model-00003-of-00004.safetensors"
965
+ }
966
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19e773648cb4e65de8660ea6365e10acca112d42a854923df93db4a6f333a82d
3
+ size 20217442
tokenizer_config.json ADDED
@@ -0,0 +1,321 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "154820": {
4
+ "content": "<|endoftext|>",
5
+ "single_word": false,
6
+ "lstrip": false,
7
+ "rstrip": false,
8
+ "normalized": false,
9
+ "special": true
10
+ },
11
+ "154821": {
12
+ "content": "[MASK]",
13
+ "single_word": false,
14
+ "lstrip": false,
15
+ "rstrip": false,
16
+ "normalized": false,
17
+ "special": true
18
+ },
19
+ "154822": {
20
+ "content": "[gMASK]",
21
+ "single_word": false,
22
+ "lstrip": false,
23
+ "rstrip": false,
24
+ "normalized": false,
25
+ "special": true
26
+ },
27
+ "154823": {
28
+ "content": "[sMASK]",
29
+ "single_word": false,
30
+ "lstrip": false,
31
+ "rstrip": false,
32
+ "normalized": false,
33
+ "special": true
34
+ },
35
+ "154824": {
36
+ "content": "<sop>",
37
+ "single_word": false,
38
+ "lstrip": false,
39
+ "rstrip": false,
40
+ "normalized": false,
41
+ "special": true
42
+ },
43
+ "154825": {
44
+ "content": "<eop>",
45
+ "single_word": false,
46
+ "lstrip": false,
47
+ "rstrip": false,
48
+ "normalized": false,
49
+ "special": true
50
+ },
51
+ "154826": {
52
+ "content": "<|system|>",
53
+ "single_word": false,
54
+ "lstrip": false,
55
+ "rstrip": false,
56
+ "normalized": false,
57
+ "special": true
58
+ },
59
+ "154827": {
60
+ "content": "<|user|>",
61
+ "single_word": false,
62
+ "lstrip": false,
63
+ "rstrip": false,
64
+ "normalized": false,
65
+ "special": true
66
+ },
67
+ "154828": {
68
+ "content": "<|assistant|>",
69
+ "single_word": false,
70
+ "lstrip": false,
71
+ "rstrip": false,
72
+ "normalized": false,
73
+ "special": true
74
+ },
75
+ "154829": {
76
+ "content": "<|observation|>",
77
+ "single_word": false,
78
+ "lstrip": false,
79
+ "rstrip": false,
80
+ "normalized": false,
81
+ "special": true
82
+ },
83
+ "154830": {
84
+ "content": "<|begin_of_image|>",
85
+ "single_word": false,
86
+ "lstrip": false,
87
+ "rstrip": false,
88
+ "normalized": false,
89
+ "special": true
90
+ },
91
+ "154831": {
92
+ "content": "<|end_of_image|>",
93
+ "single_word": false,
94
+ "lstrip": false,
95
+ "rstrip": false,
96
+ "normalized": false,
97
+ "special": true
98
+ },
99
+ "154832": {
100
+ "content": "<|begin_of_video|>",
101
+ "single_word": false,
102
+ "lstrip": false,
103
+ "rstrip": false,
104
+ "normalized": false,
105
+ "special": true
106
+ },
107
+ "154833": {
108
+ "content": "<|end_of_video|>",
109
+ "single_word": false,
110
+ "lstrip": false,
111
+ "rstrip": false,
112
+ "normalized": false,
113
+ "special": true
114
+ },
115
+ "154834": {
116
+ "content": "<|begin_of_audio|>",
117
+ "single_word": false,
118
+ "lstrip": false,
119
+ "rstrip": false,
120
+ "normalized": false,
121
+ "special": true
122
+ },
123
+ "154835": {
124
+ "content": "<|end_of_audio|>",
125
+ "single_word": false,
126
+ "lstrip": false,
127
+ "rstrip": false,
128
+ "normalized": false,
129
+ "special": true
130
+ },
131
+ "154836": {
132
+ "content": "<|begin_of_transcription|>",
133
+ "single_word": false,
134
+ "lstrip": false,
135
+ "rstrip": false,
136
+ "normalized": false,
137
+ "special": true
138
+ },
139
+ "154837": {
140
+ "content": "<|end_of_transcription|>",
141
+ "single_word": false,
142
+ "lstrip": false,
143
+ "rstrip": false,
144
+ "normalized": false,
145
+ "special": true
146
+ },
147
+ "154838": {
148
+ "content": "<|code_prefix|>",
149
+ "single_word": false,
150
+ "lstrip": false,
151
+ "rstrip": false,
152
+ "normalized": false,
153
+ "special": false
154
+ },
155
+ "154839": {
156
+ "content": "<|code_middle|>",
157
+ "single_word": false,
158
+ "lstrip": false,
159
+ "rstrip": false,
160
+ "normalized": false,
161
+ "special": false
162
+ },
163
+ "154840": {
164
+ "content": "<|code_suffix|>",
165
+ "single_word": false,
166
+ "lstrip": false,
167
+ "rstrip": false,
168
+ "normalized": false,
169
+ "special": false
170
+ },
171
+ "154841": {
172
+ "content": "<think>",
173
+ "single_word": false,
174
+ "lstrip": false,
175
+ "rstrip": false,
176
+ "normalized": false,
177
+ "special": false
178
+ },
179
+ "154842": {
180
+ "content": "</think>",
181
+ "single_word": false,
182
+ "lstrip": false,
183
+ "rstrip": false,
184
+ "normalized": false,
185
+ "special": false
186
+ },
187
+ "154843": {
188
+ "content": "<tool_call>",
189
+ "single_word": false,
190
+ "lstrip": false,
191
+ "rstrip": false,
192
+ "normalized": false,
193
+ "special": false
194
+ },
195
+ "154844": {
196
+ "content": "</tool_call>",
197
+ "single_word": false,
198
+ "lstrip": false,
199
+ "rstrip": false,
200
+ "normalized": false,
201
+ "special": false
202
+ },
203
+ "154845": {
204
+ "content": "<tool_response>",
205
+ "single_word": false,
206
+ "lstrip": false,
207
+ "rstrip": false,
208
+ "normalized": false,
209
+ "special": false
210
+ },
211
+ "154846": {
212
+ "content": "</tool_response>",
213
+ "single_word": false,
214
+ "lstrip": false,
215
+ "rstrip": false,
216
+ "normalized": false,
217
+ "special": false
218
+ },
219
+ "154847": {
220
+ "content": "<arg_key>",
221
+ "single_word": false,
222
+ "lstrip": false,
223
+ "rstrip": false,
224
+ "normalized": false,
225
+ "special": false
226
+ },
227
+ "154848": {
228
+ "content": "</arg_key>",
229
+ "single_word": false,
230
+ "lstrip": false,
231
+ "rstrip": false,
232
+ "normalized": false,
233
+ "special": false
234
+ },
235
+ "154849": {
236
+ "content": "<arg_value>",
237
+ "single_word": false,
238
+ "lstrip": false,
239
+ "rstrip": false,
240
+ "normalized": false,
241
+ "special": false
242
+ },
243
+ "154850": {
244
+ "content": "</arg_value>",
245
+ "single_word": false,
246
+ "lstrip": false,
247
+ "rstrip": false,
248
+ "normalized": false,
249
+ "special": false
250
+ },
251
+ "154851": {
252
+ "content": "/nothink",
253
+ "single_word": false,
254
+ "lstrip": false,
255
+ "rstrip": false,
256
+ "normalized": false,
257
+ "special": false
258
+ },
259
+ "154852": {
260
+ "content": "<|begin_of_box|>",
261
+ "single_word": false,
262
+ "lstrip": false,
263
+ "rstrip": false,
264
+ "normalized": false,
265
+ "special": false
266
+ },
267
+ "154853": {
268
+ "content": "<|end_of_box|>",
269
+ "single_word": false,
270
+ "lstrip": false,
271
+ "rstrip": false,
272
+ "normalized": false,
273
+ "special": false
274
+ },
275
+ "154854": {
276
+ "content": "<|image|>",
277
+ "single_word": false,
278
+ "lstrip": false,
279
+ "rstrip": false,
280
+ "normalized": false,
281
+ "special": false
282
+ },
283
+ "154855": {
284
+ "content": "<|video|>",
285
+ "single_word": false,
286
+ "lstrip": false,
287
+ "rstrip": false,
288
+ "normalized": false,
289
+ "special": false
290
+ }
291
+ },
292
+ "additional_special_tokens": [
293
+ "<|endoftext|>",
294
+ "[MASK]",
295
+ "[gMASK]",
296
+ "[sMASK]",
297
+ "<sop>",
298
+ "<eop>",
299
+ "<|system|>",
300
+ "<|user|>",
301
+ "<|assistant|>",
302
+ "<|observation|>",
303
+ "<|begin_of_image|>",
304
+ "<|end_of_image|>",
305
+ "<|begin_of_video|>",
306
+ "<|end_of_video|>",
307
+ "<|begin_of_audio|>",
308
+ "<|end_of_audio|>",
309
+ "<|begin_of_transcription|>",
310
+ "<|end_of_transcription|>"
311
+ ],
312
+ "clean_up_tokenization_spaces": false,
313
+ "do_lower_case": false,
314
+ "eos_token": "<|endoftext|>",
315
+ "extra_special_tokens": {},
316
+ "model_max_length": 128000,
317
+ "pad_token": "<|endoftext|>",
318
+ "padding_side": "left",
319
+ "remove_space": false,
320
+ "tokenizer_class": "PreTrainedTokenizer"
321
+ }