unigilby commited on
Commit
ff465b2
·
verified ·
1 Parent(s): c3882a1

Add files using upload-large-folder tool

Browse files
README.md ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: mlx
3
+ license: apache-2.0
4
+ base_model: tencent/Hy3
5
+ base_model_relation: quantized
6
+ pipeline_tag: text-generation
7
+ tags:
8
+ - mlx
9
+ - hunyuan
10
+ - hy3
11
+ - moe
12
+ - text-generation
13
+ - imatrix
14
+ ---
15
+
16
+ # Hy3-oQ4e
17
+
18
+ oMLX **oQ4e** (imatrix-enhanced affine 4-bit, group size 64, sensitivity-driven mixed
19
+ placements) quantization of [tencent/Hy3](https://huggingface.co/tencent/Hy3)
20
+ (HunYuan V3, 80-layer MoE: 192 routed experts + 1 shared, sigmoid router with expert bias).
21
+
22
+ - **158 GB** (from the 598 GB BF16 source) across 32 safetensors shards.
23
+ - **imatrix**: 128×512-token calibration (876 entries incl. per-expert statistics for all
24
+ 192 experts), collected by running an 8-bit affine proxy of the model — the BF16 source
25
+ exceeds 512 GB unified memory, so calibration ran on the half-size proxy and the resulting
26
+ activation statistics were applied to the BF16→4-bit quantization.
27
+ - **Sensitivity**: data-driven per-layer measurement (uniform 4-bit disk proxy); the most
28
+ sensitive projections (final layers 76–79, early layers 2–4) receive 5-bit boosts.
29
+ - MTP (`num_nextn_predict_layers`) weights are stripped; this is a pure decoder checkpoint.
30
+ - Coherency-verified at temperature 0 (arithmetic, factual recall, code generation,
31
+ trick-question reasoning, strict format following): 6/6.
32
+
33
+ ## Requirements
34
+
35
+ `model_type: hy_v3` support is **not yet merged** into mlx-lm — it requires
36
+ [mlx-lm PR #1211](https://github.com/ml-explore/mlx-lm/pull/1211) (adds
37
+ `mlx_lm/models/hy_v3.py`, the `hy_v3` tool parsers, and thinking-tag inference).
38
+ Apply the PR (or install from its branch) before loading:
39
+
40
+ ```python
41
+ from mlx_lm import load, generate
42
+
43
+ model, tokenizer = load("unigilby/Hy3-oQ4e")
44
+ prompt = tokenizer.apply_chat_template(
45
+ [{"role": "user", "content": "Hello!"}], add_generation_prompt=True
46
+ )
47
+ print(generate(model, tokenizer, prompt=prompt, max_tokens=200))
48
+ ```
49
+
50
+ Quantized with [oMLX](https://github.com/jundot/omlx) `quantize_oq_streaming`
51
+ (enhanced/imatrix mode) on a Mac Studio M3 Ultra.
chat_template.jinja ADDED
@@ -0,0 +1,222 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {#- ----------‑‑‑ special token variables ‑‑‑---------- -#}
2
+ {%- set HYTK = ':opensource' %}
3
+ {%- set eos_token = '<|hy_eos{}|>'.format(HYTK) %}
4
+ {%- set bos_token = '<|hy_begin_of_sentence{}|>'.format(HYTK) %}
5
+ {%- set pad_token = '<|hy_pad{}|>'.format(HYTK) %}
6
+ {%- set user_token = '<|hy_User{}|>'.format(HYTK) %}
7
+ {%- set assistant_token = '<|hy_Assistant{}|>'.format(HYTK) %}
8
+ {%- set think_begin_token = '<think{}>'.format(HYTK) %}
9
+ {%- set think_end_token = '</think{}>'.format(HYTK) %}
10
+ {%- set toolcalls_begin_token = '<tool_calls{}>'.format(HYTK) %}
11
+ {%- set toolcalls_end_token = '</tool_calls{}>'.format(HYTK) %}
12
+ {%- set toolcall_begin_token = '<tool_call{}>'.format(HYTK) %}
13
+ {%- set toolcall_end_token = '</tool_call{}>'.format(HYTK) %}
14
+ {%- set toolsep_token = '<tool_sep{}>'.format(HYTK) %}
15
+ {%- set argkey_begin_token = '<arg_key{}>'.format(HYTK) %}
16
+ {%- set argkey_end_token = '</arg_key{}>'.format(HYTK) %}
17
+ {%- set argvalue_begin_token = '<arg_value{}>'.format(HYTK) %}
18
+ {%- set argvalue_end_token = '</arg_value{}>'.format(HYTK) %}
19
+ {%- set toolresponses_begin_token = '<tool_responses{}>'.format(HYTK) %}
20
+ {%- set toolresponses_end_token = '</tool_responses{}>'.format(HYTK) %}
21
+ {%- set toolresponse_begin_token = '<tool_response{}>'.format(HYTK) %}
22
+ {%- set toolresponse_end_token = '</tool_response{}>'.format(HYTK) %}
23
+ {%- set reasoning_mode_token = '<|reasoning_mode{}|>'.format(HYTK) %}
24
+
25
+ {#- ----------‑‑‑ hyperparameters variables ‑‑‑---------- -#}
26
+ {%- if not add_generation_prompt is defined %}
27
+ {%- set add_generation_prompt = false %}
28
+ {%- endif %}
29
+ {%- if not preserved_thinking is defined %}
30
+ {%- if not tools %}
31
+ {%- set preserved_thinking = false %}
32
+ {%- else %}
33
+ {%- set preserved_thinking = true %}
34
+ {%- endif %}
35
+ {%- endif %}
36
+ {%- if not is_training is defined %}
37
+ {%- set is_training = false %}
38
+ {%- endif %}
39
+
40
+ {%- if not reasoning_effort is defined %}
41
+ {%- set reasoning_effort = 'no_think' %}
42
+ {%- elif reasoning_effort not in ['high', 'low', 'no_think'] %}
43
+ {%- if reasoning_effort is none %}
44
+ {{- raise_exception('reasoning_effort error : None, should be no_think/low/high') }}
45
+ {%- else %}
46
+ {{- raise_exception('reasoning_effort error : ' + reasoning_effort + ', should be no_think/low/high') }}
47
+ {%- endif %}
48
+ {%- endif %}
49
+
50
+ {%- if fallback_strategy is defined and fallback_strategy == 'reasoning_toolcall_retry' %}
51
+ {%- set reasoning_effort = 'high' %}
52
+ {%- set add_generation_prompt = false %}
53
+ {%- endif %}
54
+ {%- if not raw_last_assistant is defined %}
55
+ {%- set raw_last_assistant = false %}
56
+ {%- endif %}
57
+
58
+ {%- macro visible_text(content) -%}
59
+ {%- if content is string -%}
60
+ {{- content }}
61
+ {%- elif content is iterable and content is not mapping -%}
62
+ {%- for item in content -%}
63
+ {%- if item is mapping and item.type == 'text' -%}
64
+ {{- item.text }}
65
+ {%- elif item is string -%}
66
+ {{- item }}
67
+ {%- endif -%}
68
+ {%- endfor -%}
69
+ {%- elif content is none -%}
70
+ {{- '' }}
71
+ {%- else -%}
72
+ {{- content }}
73
+ {%- endif -%}
74
+ {%- endmacro -%}
75
+
76
+ {%- set ns = namespace(last_user_index=-1) %}
77
+ {%- set sp_ns = namespace(system_prompt='', is_first_sp=true) %}
78
+ {%- for message in messages %}
79
+ {%- if message['role'] == 'system' %}
80
+ {%- set sp_ns.system_prompt = sp_ns.system_prompt + visible_text(message['content']) %}
81
+ {%- endif %}
82
+ {%- if message['role'] == 'user' %}
83
+ {%- set ns.last_user_index = loop.index0 %}
84
+ {%- endif %}
85
+ {%- endfor %}
86
+ {%- if reasoning_effort is defined and reasoning_effort is string and reasoning_effort != '' and not tools %}
87
+ {%- set sp_ns.system_prompt = sp_ns.system_prompt + reasoning_mode_token + 'reasoning_effort:' + reasoning_effort %}
88
+ {%- endif %}
89
+ {{- bos_token }}
90
+ {{- sp_ns.system_prompt }}
91
+ {%- if tools %}
92
+ {%- if sp_ns.system_prompt != '' %}
93
+ {{- '\n\n# Tools\n\nYou may call one or more functions to assist with the user query.' }}
94
+ {%- else %}
95
+ {{- '# Tools\n\nYou may call one or more functions to assist with the user query.' }}
96
+ {%- endif %}
97
+ {{- '\n\nYou are provided with function signatures within <tools></tools> XML tags:' }}
98
+ {{- '\n<tools>\n' }}
99
+ {%- for tool in tools %}
100
+ {%- if loop.index0 > 0 %}
101
+ {{- '\n' }}
102
+ {%- endif %}
103
+ {{- tool | tojson }}
104
+ {%- endfor %}
105
+ {{- '\n</tools>\n\n' }}
106
+ {{- 'For function call returns, you should first print ' + toolcalls_begin_token + '\n' }}
107
+ {{- 'For each function call, you should return object like:\n' }}
108
+ {{- toolcall_begin_token + '{function-name}' + toolsep_token + '\n' }}
109
+ {{- argkey_begin_token + '{arg-key-1}' + argkey_end_token + '\n' }}
110
+ {{- argvalue_begin_token + '{arg-value-1}' + argvalue_end_token + '\n' }}
111
+ {{- argkey_begin_token + '{arg-key-2}' + argkey_end_token + '\n' }}
112
+ {{- argvalue_begin_token + '{arg-value-2}' + argvalue_end_token + '\n' }}
113
+ {{- '...\n' }}
114
+ {{- toolcall_end_token + '\n' }}
115
+ {%- if reasoning_effort is defined and reasoning_effort is string and reasoning_effort != '' %}
116
+ {{- 'At the end of function call returns, you should print ' + toolcalls_end_token + reasoning_mode_token + 'reasoning_effort:' + reasoning_effort }}
117
+ {%- else %}
118
+ {{- 'At the end of function call returns, you should print ' + toolcalls_end_token }}
119
+ {%- endif %}
120
+ {%- endif %}
121
+
122
+ {%- set prev_ns = namespace(is_tool=false, is_tool_first=true) %}
123
+ {%- set last_ns = namespace(last_is_assistant=false) %}
124
+ {%- for message in messages %}
125
+ {%- if message['role'] == 'user' %}
126
+ {%- if prev_ns.is_tool %}
127
+ {{- toolresponses_end_token }}
128
+ {%- endif %}
129
+ {{- user_token + visible_text(message['content']) }}
130
+ {%- set prev_ns.is_tool = false %}
131
+ {%- endif %}
132
+ {%- if message['role'] == 'assistant' %}
133
+ {%- if is_training %}
134
+ {%- if 'reasoning_content' in message and message['reasoning_content'] is string %}
135
+ {%- set rc = message['reasoning_content'] %}
136
+ {%- elif 'reasoning' in message and message['reasoning'] is string %}
137
+ {%- set rc = message['reasoning'] %}
138
+ {%- else %}
139
+ {%- set rc = none %}
140
+ {%- endif %}
141
+ {%- if rc is not none %}
142
+ {%- set content = think_begin_token + rc + think_end_token + visible_text(message['content']) %}
143
+ {%- else %}
144
+ {%- set content = think_begin_token + think_end_token + visible_text(message['content']) %}
145
+ {%- endif %}
146
+ {%- else %}
147
+ {%- if ((preserved_thinking is defined and preserved_thinking) or loop.index0 > ns.last_user_index) %}
148
+ {%- if 'reasoning_content' in message and message['reasoning_content'] is string %}
149
+ {%- set rc = message['reasoning_content'] %}
150
+ {%- elif 'reasoning' in message and message['reasoning'] is string %}
151
+ {%- set rc = message['reasoning'] %}
152
+ {%- else %}
153
+ {%- set rc = none %}
154
+ {%- endif %}
155
+ {%- if rc is not none %}
156
+ {%- set content = think_begin_token + rc + think_end_token + visible_text(message['content']) %}
157
+ {%- else %}
158
+ {%- set content = think_begin_token + think_end_token + visible_text(message['content']) %}
159
+ {%- endif %}
160
+ {%- else %}
161
+ {%- set content = think_begin_token + think_end_token + visible_text(message['content']) %}
162
+ {%- endif %}
163
+ {%- endif %}
164
+ {%- if prev_ns.is_tool %}
165
+ {{- toolresponses_end_token }}
166
+ {%- endif %}
167
+ {{- assistant_token }}
168
+ {%- if message['tool_calls'] is defined and message['tool_calls'] %}
169
+ {%- set prev_ns.is_tool_first = true %}
170
+ {{- content }}
171
+ {{- toolcalls_begin_token + '\n' }}
172
+ {%- for tool in message['tool_calls'] %}
173
+ {%- set arguments = tool['function']['arguments'] %}
174
+ {{- toolcall_begin_token + tool['function']['name'] + toolsep_token + '\n' }}
175
+ {%- for key, value in arguments.items() %}
176
+ {{- argkey_begin_token + key + argkey_end_token + '\n' }}
177
+ {%- if value is not string %}
178
+ {%- set value = value | tojson(ensure_ascii=False) %}
179
+ {%- endif %}
180
+ {{- argvalue_begin_token + value + argvalue_end_token + '\n' }}
181
+ {%- endfor %}
182
+ {{- toolcall_end_token + '\n' }}
183
+ {%- endfor %}
184
+ {{- toolcalls_end_token + eos_token }}
185
+ {%- else %}
186
+ {%- if loop.last and raw_last_assistant %}
187
+ {{- visible_text(message['content']) }}
188
+ {%- elif not loop.last or is_training %}
189
+ {{- content + eos_token }}
190
+ {%- else %}
191
+ {{- content }}
192
+ {%- endif %}
193
+ {%- endif %}
194
+ {%- set prev_ns.is_tool = false %}
195
+ {%- endif %}
196
+ {%- if message['role'] == 'tool' %}
197
+ {%- set prev_ns.is_tool = true %}
198
+ {%- if prev_ns.is_tool_first %}
199
+ {{- toolresponses_begin_token + '\n' }}
200
+ {%- set prev_ns.is_tool_first = false %}
201
+ {%- endif %}
202
+ {{- toolresponse_begin_token + '\n' + visible_text(message['content']) + '\n' + toolresponse_end_token + '\n' }}
203
+ {%- endif %}
204
+ {%- if loop.last and message['role'] == 'assistant' %}
205
+ {%- set last_ns.last_is_assistant = true %}
206
+ {%- endif %}
207
+
208
+ {%- endfor %}
209
+ {%- if prev_ns.is_tool %}
210
+ {{- toolresponses_end_token }}
211
+ {%- endif %}
212
+ {%- if add_generation_prompt %}
213
+ {%- if not last_ns.last_is_assistant %}
214
+ {%- if reasoning_effort is defined and reasoning_effort in ['low', 'high'] %}
215
+ {{- assistant_token + think_begin_token }}
216
+ {%- elif reasoning_effort is defined and reasoning_effort == 'no_think' %}
217
+ {{- assistant_token + think_begin_token + think_end_token }}
218
+ {%- else %}
219
+ {{- assistant_token }}
220
+ {%- endif %}
221
+ {%- endif %}
222
+ {%- endif %}
config.json ADDED
The diff for this file is too large to render. See raw diff
 
generation_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 120000,
3
+ "do_sample": true,
4
+ "eos_token_id": 120025,
5
+ "pad_token_id": 120002,
6
+ "temperature": 0.9,
7
+ "top_k": -1,
8
+ "top_p": 1,
9
+ "transformers_version": "5.6.0"
10
+ }
model-00001-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2104cef5e9f8320def1d3618ed083fa6e936dc6eeb27a0a6c00dee330836ded5
3
+ size 5029416223
model-00002-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:45d8cc10bd8eb69c39784e3efc7c5760105ca23b58aa3dabd13e787e85fb88f4
3
+ size 5078498037
model-00003-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dee2f66bb1c4194f35ac771673a294c453c38814070a9ea04c6ae75cde433750
3
+ size 5435823661
model-00004-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:66c609ff6c549f846b5573ad6ccafca89e18393c2aacd7f981cc781abac5c1e3
3
+ size 5435822792
model-00005-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bcf6aacfcf0fb9480801a221886337bf24f60a9026cfc1e9fab32a3319bc5e12
3
+ size 5435823665
model-00006-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c6e7538bd6e6d8bef4e7b8d2dc0901d7bbcd110390b9182df6ed2cb260cd02d5
3
+ size 5435823679
model-00007-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:edbe950e4abec41dc2a46109dbc1d7093f69f6a16746455bec76bed9fcd8798f
3
+ size 5435822820
model-00008-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6350a8434b1f6823d3b31e15677fbda893af5f6524794c2af6e5f9f02e28500c
3
+ size 5435823694
model-00009-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6167dd1fd263df281e6f5dd5a369a3107d6aed789ccdf6cad03ba84a909f8fff
3
+ size 5435823688
model-00010-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d98fc38e2453e377bc143f5c8e7c5ab003751e57f60b0837012befb70433740c
3
+ size 5435822816
model-00011-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:09bcc4a6ab4a3f4a9b6c97630dcf6a29503f27230116a62eab0097d0de89843a
3
+ size 5435823692
model-00012-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a00ac15e4feac148c7add1f9c83980bdcd1c5a744901c74396892102488ac031
3
+ size 5435823686
model-00013-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f137230d486f1dccf398806664672e4b7f24cd680a6aae01c6049df316416437
3
+ size 5435822806
model-00014-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5c146d3629fed211e46672c6763160ca4f00f8fa3b352babfcf6bb2f9df08f2a
3
+ size 5435823692
model-00015-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:163ca55e42e5ecd74a44f03797cc598a1ccac768fec1e5b39f40d756eb9f0598
3
+ size 5435823686
model-00016-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ce6b9e1d56e5f1780868c59c40a2204865de774b7409f7db72ee61fba2c0a0e
3
+ size 5435822818
model-00017-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3bc406cc47a7df33b295c1c9688a1e21ecda9efeac852293486c7997d2d92d26
3
+ size 5435823692
model-00018-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:36517cf8709e787747eb652ad164c2d308a1f4e5eecb80b5d9ecffb4942b2f1b
3
+ size 5435823680
model-00019-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:150f4bbb87694398e9dd0efc62cf6da0c2bd785987e43aab4b6f45ea3b4f9937
3
+ size 5435822814
model-00020-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b9cfc8cbe9db5e5871e467a5e63d163cbf264bc4b873de3d4a615e16939d2c1e
3
+ size 5435823698
model-00021-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b4e9aa43c6c6071de45fb21501b075994d0106edffb96e5986d699562d38ef14
3
+ size 5435823684
model-00022-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e40970dac9ab62542e2e4887de9371a2392ea7c797cacaa530791c019d6397c
3
+ size 5435822812
model-00023-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ed1970e6eaf6cbd49a8246ef2559b40f0abe0eabd1f6e3cee0dd92ef52a16c5c
3
+ size 5435823700
model-00024-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f59dc4626b6422afcc8f8d0cadb0a6dc66255ac6276aec9b6e168f55f4ad5800
3
+ size 5435823690
model-00025-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:569196b36d54d4f782f7a4bb8808be786951719350c729f184c462427f87bc56
3
+ size 5435822814
model-00026-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6c41bb9e2278e398cc42b0b16ff09f85621338afab4abe6d4e86f844fb4bb37f
3
+ size 5435823700
model-00027-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:178d808eeb3b351d0f85cc591ea85a05692fb6343d85b7603b1ae608a14d9c68
3
+ size 5435823682
model-00028-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d9458f14561e4612292b8fe374f3ddee1ece258a166a940896c98e463a9cd9d0
3
+ size 5435822800
model-00029-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:72c8ae0d51282cad3333d713c466e68aef73f37e462f90ca4f3b7cb4e705016e
3
+ size 5435823682
model-00030-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d0b7402c31cd0a67bd77ecd5181b77ded43ed8c19c084f76e8088a1f542fcb65
3
+ size 5435823678
model-00031-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad002baef82d869e566ed6ae022ff5f0fb58c80a65bb1084df4e19bc0efc2e4c
3
+ size 5435822818
model-00032-of-00032.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4055448acb9e869420e0aa29dafa3d372bb64aafa303f14a50513c553001ccf2
3
+ size 2038433782
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
oq_imatrix_report.json ADDED
@@ -0,0 +1,888 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "enabled": true,
3
+ "cache_path": "/Users/earl/models/tencent/.oqe_imatrix/Hy3-oQe-s128-l512.npz",
4
+ "cache_reused": true,
5
+ "entry_count": 876,
6
+ "calib_dataset": "oqe_code_multilingual",
7
+ "collection": {
8
+ "dataset": "oqe_code_multilingual",
9
+ "requested_samples": 128,
10
+ "seq_length": 512,
11
+ "adaptive": true,
12
+ "adaptive_step_samples": 128,
13
+ "adaptive_max_samples": 1024,
14
+ "available_samples": 1024,
15
+ "micro_batch_size": 12,
16
+ "micro_batches": 11,
17
+ "batch_plan": {
18
+ "micro_batch_size": 12,
19
+ "estimated_sample_bytes": 67108864,
20
+ "capture_budget_bytes": 805306368,
21
+ "system_available_bytes": 539079163904,
22
+ "metal_available_bytes": 515391651504,
23
+ "live_available_bytes": 515391651504,
24
+ "hidden_size": 4096,
25
+ "num_experts": 192,
26
+ "top_k": 8
27
+ },
28
+ "processed_samples": 128,
29
+ "installed_modules": 877,
30
+ "coverage_sufficient": true,
31
+ "coverage": {
32
+ "has_expert_counts": true,
33
+ "expert_modules": 237,
34
+ "total_experts": 45504,
35
+ "active_experts": 45504,
36
+ "zero_count_experts": 0,
37
+ "active_ratio": 1.0,
38
+ "min_count": 25,
39
+ "p05_count": 1065.0,
40
+ "p10_count": 1327.0,
41
+ "median_count": 2462.0,
42
+ "max_count": 28519,
43
+ "min_required_count": 16,
44
+ "required_percentile": 5
45
+ },
46
+ "rounds": [
47
+ {
48
+ "processed_samples": 128,
49
+ "coverage_sufficient": true,
50
+ "coverage": {
51
+ "has_expert_counts": true,
52
+ "expert_modules": 237,
53
+ "total_experts": 45504,
54
+ "active_experts": 45504,
55
+ "zero_count_experts": 0,
56
+ "active_ratio": 1.0,
57
+ "min_count": 25,
58
+ "p05_count": 1065.0,
59
+ "p10_count": 1327.0,
60
+ "median_count": 2462.0,
61
+ "max_count": 28519,
62
+ "min_required_count": 16,
63
+ "required_percentile": 5
64
+ }
65
+ }
66
+ ]
67
+ },
68
+ "expert_coverage": {
69
+ "has_expert_counts": true,
70
+ "expert_modules": 237,
71
+ "total_experts": 45504,
72
+ "active_experts": 45504,
73
+ "zero_count_experts": 0,
74
+ "active_ratio": 1.0,
75
+ "min_count": 25,
76
+ "p05_count": 1065.0,
77
+ "p10_count": 1327.0,
78
+ "median_count": 2462.0,
79
+ "max_count": 28519,
80
+ "min_required_count": 16,
81
+ "required_percentile": 5
82
+ },
83
+ "applied": [
84
+ "model.layers.0.mlp.down_proj",
85
+ "model.layers.0.mlp.gate_proj",
86
+ "model.layers.0.mlp.up_proj",
87
+ "model.layers.0.self_attn.k_proj",
88
+ "model.layers.0.self_attn.o_proj",
89
+ "model.layers.0.self_attn.q_proj",
90
+ "model.layers.0.self_attn.v_proj",
91
+ "model.layers.1.mlp.shared_mlp.down_proj",
92
+ "model.layers.1.mlp.shared_mlp.gate_proj",
93
+ "model.layers.1.mlp.shared_mlp.up_proj",
94
+ "model.layers.1.mlp.switch_mlp.down_proj",
95
+ "model.layers.1.mlp.switch_mlp.gate_proj",
96
+ "model.layers.1.mlp.switch_mlp.up_proj",
97
+ "model.layers.1.self_attn.k_proj",
98
+ "model.layers.1.self_attn.o_proj",
99
+ "model.layers.1.self_attn.q_proj",
100
+ "model.layers.1.self_attn.v_proj",
101
+ "model.layers.10.mlp.shared_mlp.down_proj",
102
+ "model.layers.10.mlp.shared_mlp.gate_proj",
103
+ "model.layers.10.mlp.shared_mlp.up_proj",
104
+ "model.layers.10.mlp.switch_mlp.down_proj",
105
+ "model.layers.10.mlp.switch_mlp.gate_proj",
106
+ "model.layers.10.mlp.switch_mlp.up_proj",
107
+ "model.layers.10.self_attn.k_proj",
108
+ "model.layers.10.self_attn.o_proj",
109
+ "model.layers.10.self_attn.q_proj",
110
+ "model.layers.10.self_attn.v_proj",
111
+ "model.layers.11.mlp.shared_mlp.down_proj",
112
+ "model.layers.11.mlp.shared_mlp.gate_proj",
113
+ "model.layers.11.mlp.shared_mlp.up_proj",
114
+ "model.layers.11.mlp.switch_mlp.down_proj",
115
+ "model.layers.11.mlp.switch_mlp.gate_proj",
116
+ "model.layers.11.mlp.switch_mlp.up_proj",
117
+ "model.layers.11.self_attn.k_proj",
118
+ "model.layers.11.self_attn.o_proj",
119
+ "model.layers.11.self_attn.q_proj",
120
+ "model.layers.11.self_attn.v_proj",
121
+ "model.layers.12.mlp.shared_mlp.down_proj",
122
+ "model.layers.12.mlp.shared_mlp.gate_proj",
123
+ "model.layers.12.mlp.shared_mlp.up_proj",
124
+ "model.layers.12.mlp.switch_mlp.down_proj",
125
+ "model.layers.12.mlp.switch_mlp.gate_proj",
126
+ "model.layers.12.mlp.switch_mlp.up_proj",
127
+ "model.layers.12.self_attn.k_proj",
128
+ "model.layers.12.self_attn.o_proj",
129
+ "model.layers.12.self_attn.q_proj",
130
+ "model.layers.12.self_attn.v_proj",
131
+ "model.layers.13.mlp.shared_mlp.down_proj",
132
+ "model.layers.13.mlp.shared_mlp.gate_proj",
133
+ "model.layers.13.mlp.shared_mlp.up_proj",
134
+ "model.layers.13.mlp.switch_mlp.down_proj",
135
+ "model.layers.13.mlp.switch_mlp.gate_proj",
136
+ "model.layers.13.mlp.switch_mlp.up_proj",
137
+ "model.layers.13.self_attn.k_proj",
138
+ "model.layers.13.self_attn.o_proj",
139
+ "model.layers.13.self_attn.q_proj",
140
+ "model.layers.13.self_attn.v_proj",
141
+ "model.layers.14.mlp.shared_mlp.down_proj",
142
+ "model.layers.14.mlp.shared_mlp.gate_proj",
143
+ "model.layers.14.mlp.shared_mlp.up_proj",
144
+ "model.layers.14.mlp.switch_mlp.down_proj",
145
+ "model.layers.14.mlp.switch_mlp.gate_proj",
146
+ "model.layers.14.mlp.switch_mlp.up_proj",
147
+ "model.layers.14.self_attn.k_proj",
148
+ "model.layers.14.self_attn.o_proj",
149
+ "model.layers.14.self_attn.q_proj",
150
+ "model.layers.14.self_attn.v_proj",
151
+ "model.layers.15.mlp.shared_mlp.down_proj",
152
+ "model.layers.15.mlp.shared_mlp.gate_proj",
153
+ "model.layers.15.mlp.shared_mlp.up_proj",
154
+ "model.layers.15.mlp.switch_mlp.down_proj",
155
+ "model.layers.15.mlp.switch_mlp.gate_proj",
156
+ "model.layers.15.mlp.switch_mlp.up_proj",
157
+ "model.layers.15.self_attn.k_proj",
158
+ "model.layers.15.self_attn.o_proj",
159
+ "model.layers.15.self_attn.q_proj",
160
+ "model.layers.15.self_attn.v_proj",
161
+ "model.layers.16.mlp.shared_mlp.down_proj",
162
+ "model.layers.16.mlp.shared_mlp.gate_proj",
163
+ "model.layers.16.mlp.shared_mlp.up_proj",
164
+ "model.layers.16.mlp.switch_mlp.down_proj",
165
+ "model.layers.16.mlp.switch_mlp.gate_proj",
166
+ "model.layers.16.mlp.switch_mlp.up_proj",
167
+ "model.layers.16.self_attn.k_proj",
168
+ "model.layers.16.self_attn.o_proj",
169
+ "model.layers.16.self_attn.q_proj",
170
+ "model.layers.16.self_attn.v_proj",
171
+ "model.layers.17.mlp.shared_mlp.down_proj",
172
+ "model.layers.17.mlp.shared_mlp.gate_proj",
173
+ "model.layers.17.mlp.shared_mlp.up_proj",
174
+ "model.layers.17.mlp.switch_mlp.down_proj",
175
+ "model.layers.17.mlp.switch_mlp.gate_proj",
176
+ "model.layers.17.mlp.switch_mlp.up_proj",
177
+ "model.layers.17.self_attn.k_proj",
178
+ "model.layers.17.self_attn.o_proj",
179
+ "model.layers.17.self_attn.q_proj",
180
+ "model.layers.17.self_attn.v_proj",
181
+ "model.layers.18.mlp.shared_mlp.down_proj",
182
+ "model.layers.18.mlp.shared_mlp.gate_proj",
183
+ "model.layers.18.mlp.shared_mlp.up_proj",
184
+ "model.layers.18.mlp.switch_mlp.down_proj",
185
+ "model.layers.18.mlp.switch_mlp.gate_proj",
186
+ "model.layers.18.mlp.switch_mlp.up_proj",
187
+ "model.layers.18.self_attn.k_proj",
188
+ "model.layers.18.self_attn.o_proj",
189
+ "model.layers.18.self_attn.q_proj",
190
+ "model.layers.18.self_attn.v_proj",
191
+ "model.layers.19.mlp.shared_mlp.down_proj",
192
+ "model.layers.19.mlp.shared_mlp.gate_proj",
193
+ "model.layers.19.mlp.shared_mlp.up_proj",
194
+ "model.layers.19.mlp.switch_mlp.down_proj",
195
+ "model.layers.19.mlp.switch_mlp.gate_proj",
196
+ "model.layers.19.mlp.switch_mlp.up_proj",
197
+ "model.layers.19.self_attn.k_proj",
198
+ "model.layers.19.self_attn.o_proj",
199
+ "model.layers.19.self_attn.q_proj",
200
+ "model.layers.19.self_attn.v_proj",
201
+ "model.layers.2.mlp.shared_mlp.down_proj",
202
+ "model.layers.2.mlp.shared_mlp.gate_proj",
203
+ "model.layers.2.mlp.shared_mlp.up_proj",
204
+ "model.layers.2.mlp.switch_mlp.down_proj",
205
+ "model.layers.2.mlp.switch_mlp.gate_proj",
206
+ "model.layers.2.mlp.switch_mlp.up_proj",
207
+ "model.layers.2.self_attn.k_proj",
208
+ "model.layers.2.self_attn.o_proj",
209
+ "model.layers.2.self_attn.q_proj",
210
+ "model.layers.2.self_attn.v_proj",
211
+ "model.layers.20.mlp.shared_mlp.down_proj",
212
+ "model.layers.20.mlp.shared_mlp.gate_proj",
213
+ "model.layers.20.mlp.shared_mlp.up_proj",
214
+ "model.layers.20.mlp.switch_mlp.down_proj",
215
+ "model.layers.20.mlp.switch_mlp.gate_proj",
216
+ "model.layers.20.mlp.switch_mlp.up_proj",
217
+ "model.layers.20.self_attn.k_proj",
218
+ "model.layers.20.self_attn.o_proj",
219
+ "model.layers.20.self_attn.q_proj",
220
+ "model.layers.20.self_attn.v_proj",
221
+ "model.layers.21.mlp.shared_mlp.down_proj",
222
+ "model.layers.21.mlp.shared_mlp.gate_proj",
223
+ "model.layers.21.mlp.shared_mlp.up_proj",
224
+ "model.layers.21.mlp.switch_mlp.down_proj",
225
+ "model.layers.21.mlp.switch_mlp.gate_proj",
226
+ "model.layers.21.mlp.switch_mlp.up_proj",
227
+ "model.layers.21.self_attn.k_proj",
228
+ "model.layers.21.self_attn.o_proj",
229
+ "model.layers.21.self_attn.q_proj",
230
+ "model.layers.21.self_attn.v_proj",
231
+ "model.layers.22.mlp.shared_mlp.down_proj",
232
+ "model.layers.22.mlp.shared_mlp.gate_proj",
233
+ "model.layers.22.mlp.shared_mlp.up_proj",
234
+ "model.layers.22.mlp.switch_mlp.down_proj",
235
+ "model.layers.22.mlp.switch_mlp.gate_proj",
236
+ "model.layers.22.mlp.switch_mlp.up_proj",
237
+ "model.layers.22.self_attn.k_proj",
238
+ "model.layers.22.self_attn.o_proj",
239
+ "model.layers.22.self_attn.q_proj",
240
+ "model.layers.22.self_attn.v_proj",
241
+ "model.layers.23.mlp.shared_mlp.down_proj",
242
+ "model.layers.23.mlp.shared_mlp.gate_proj",
243
+ "model.layers.23.mlp.shared_mlp.up_proj",
244
+ "model.layers.23.mlp.switch_mlp.down_proj",
245
+ "model.layers.23.mlp.switch_mlp.gate_proj",
246
+ "model.layers.23.mlp.switch_mlp.up_proj",
247
+ "model.layers.23.self_attn.k_proj",
248
+ "model.layers.23.self_attn.o_proj",
249
+ "model.layers.23.self_attn.q_proj",
250
+ "model.layers.23.self_attn.v_proj",
251
+ "model.layers.24.mlp.shared_mlp.down_proj",
252
+ "model.layers.24.mlp.shared_mlp.gate_proj",
253
+ "model.layers.24.mlp.shared_mlp.up_proj",
254
+ "model.layers.24.mlp.switch_mlp.down_proj",
255
+ "model.layers.24.mlp.switch_mlp.gate_proj",
256
+ "model.layers.24.mlp.switch_mlp.up_proj",
257
+ "model.layers.24.self_attn.k_proj",
258
+ "model.layers.24.self_attn.o_proj",
259
+ "model.layers.24.self_attn.q_proj",
260
+ "model.layers.24.self_attn.v_proj",
261
+ "model.layers.25.mlp.shared_mlp.down_proj",
262
+ "model.layers.25.mlp.shared_mlp.gate_proj",
263
+ "model.layers.25.mlp.shared_mlp.up_proj",
264
+ "model.layers.25.mlp.switch_mlp.down_proj",
265
+ "model.layers.25.mlp.switch_mlp.gate_proj",
266
+ "model.layers.25.mlp.switch_mlp.up_proj",
267
+ "model.layers.25.self_attn.k_proj",
268
+ "model.layers.25.self_attn.o_proj",
269
+ "model.layers.25.self_attn.q_proj",
270
+ "model.layers.25.self_attn.v_proj",
271
+ "model.layers.26.mlp.shared_mlp.down_proj",
272
+ "model.layers.26.mlp.shared_mlp.gate_proj",
273
+ "model.layers.26.mlp.shared_mlp.up_proj",
274
+ "model.layers.26.mlp.switch_mlp.down_proj",
275
+ "model.layers.26.mlp.switch_mlp.gate_proj",
276
+ "model.layers.26.mlp.switch_mlp.up_proj",
277
+ "model.layers.26.self_attn.k_proj",
278
+ "model.layers.26.self_attn.o_proj",
279
+ "model.layers.26.self_attn.q_proj",
280
+ "model.layers.26.self_attn.v_proj",
281
+ "model.layers.27.mlp.shared_mlp.down_proj",
282
+ "model.layers.27.mlp.shared_mlp.gate_proj",
283
+ "model.layers.27.mlp.shared_mlp.up_proj",
284
+ "model.layers.27.mlp.switch_mlp.down_proj",
285
+ "model.layers.27.mlp.switch_mlp.gate_proj",
286
+ "model.layers.27.mlp.switch_mlp.up_proj",
287
+ "model.layers.27.self_attn.k_proj",
288
+ "model.layers.27.self_attn.o_proj",
289
+ "model.layers.27.self_attn.q_proj",
290
+ "model.layers.27.self_attn.v_proj",
291
+ "model.layers.28.mlp.shared_mlp.down_proj",
292
+ "model.layers.28.mlp.shared_mlp.gate_proj",
293
+ "model.layers.28.mlp.shared_mlp.up_proj",
294
+ "model.layers.28.mlp.switch_mlp.down_proj",
295
+ "model.layers.28.mlp.switch_mlp.gate_proj",
296
+ "model.layers.28.mlp.switch_mlp.up_proj",
297
+ "model.layers.28.self_attn.k_proj",
298
+ "model.layers.28.self_attn.o_proj",
299
+ "model.layers.28.self_attn.q_proj",
300
+ "model.layers.28.self_attn.v_proj",
301
+ "model.layers.29.mlp.shared_mlp.down_proj",
302
+ "model.layers.29.mlp.shared_mlp.gate_proj",
303
+ "model.layers.29.mlp.shared_mlp.up_proj",
304
+ "model.layers.29.mlp.switch_mlp.down_proj",
305
+ "model.layers.29.mlp.switch_mlp.gate_proj",
306
+ "model.layers.29.mlp.switch_mlp.up_proj",
307
+ "model.layers.29.self_attn.k_proj",
308
+ "model.layers.29.self_attn.o_proj",
309
+ "model.layers.29.self_attn.q_proj",
310
+ "model.layers.29.self_attn.v_proj",
311
+ "model.layers.3.mlp.shared_mlp.down_proj",
312
+ "model.layers.3.mlp.shared_mlp.gate_proj",
313
+ "model.layers.3.mlp.shared_mlp.up_proj",
314
+ "model.layers.3.mlp.switch_mlp.down_proj",
315
+ "model.layers.3.mlp.switch_mlp.gate_proj",
316
+ "model.layers.3.mlp.switch_mlp.up_proj",
317
+ "model.layers.3.self_attn.k_proj",
318
+ "model.layers.3.self_attn.o_proj",
319
+ "model.layers.3.self_attn.q_proj",
320
+ "model.layers.3.self_attn.v_proj",
321
+ "model.layers.30.mlp.shared_mlp.down_proj",
322
+ "model.layers.30.mlp.shared_mlp.gate_proj",
323
+ "model.layers.30.mlp.shared_mlp.up_proj",
324
+ "model.layers.30.mlp.switch_mlp.down_proj",
325
+ "model.layers.30.mlp.switch_mlp.gate_proj",
326
+ "model.layers.30.mlp.switch_mlp.up_proj",
327
+ "model.layers.30.self_attn.k_proj",
328
+ "model.layers.30.self_attn.o_proj",
329
+ "model.layers.30.self_attn.q_proj",
330
+ "model.layers.30.self_attn.v_proj",
331
+ "model.layers.31.mlp.shared_mlp.down_proj",
332
+ "model.layers.31.mlp.shared_mlp.gate_proj",
333
+ "model.layers.31.mlp.shared_mlp.up_proj",
334
+ "model.layers.31.mlp.switch_mlp.down_proj",
335
+ "model.layers.31.mlp.switch_mlp.gate_proj",
336
+ "model.layers.31.mlp.switch_mlp.up_proj",
337
+ "model.layers.31.self_attn.k_proj",
338
+ "model.layers.31.self_attn.o_proj",
339
+ "model.layers.31.self_attn.q_proj",
340
+ "model.layers.31.self_attn.v_proj",
341
+ "model.layers.32.mlp.shared_mlp.down_proj",
342
+ "model.layers.32.mlp.shared_mlp.gate_proj",
343
+ "model.layers.32.mlp.shared_mlp.up_proj",
344
+ "model.layers.32.mlp.switch_mlp.down_proj",
345
+ "model.layers.32.mlp.switch_mlp.gate_proj",
346
+ "model.layers.32.mlp.switch_mlp.up_proj",
347
+ "model.layers.32.self_attn.k_proj",
348
+ "model.layers.32.self_attn.o_proj",
349
+ "model.layers.32.self_attn.q_proj",
350
+ "model.layers.32.self_attn.v_proj",
351
+ "model.layers.33.mlp.shared_mlp.down_proj",
352
+ "model.layers.33.mlp.shared_mlp.gate_proj",
353
+ "model.layers.33.mlp.shared_mlp.up_proj",
354
+ "model.layers.33.mlp.switch_mlp.down_proj",
355
+ "model.layers.33.mlp.switch_mlp.gate_proj",
356
+ "model.layers.33.mlp.switch_mlp.up_proj",
357
+ "model.layers.33.self_attn.k_proj",
358
+ "model.layers.33.self_attn.o_proj",
359
+ "model.layers.33.self_attn.q_proj",
360
+ "model.layers.33.self_attn.v_proj",
361
+ "model.layers.34.mlp.shared_mlp.down_proj",
362
+ "model.layers.34.mlp.shared_mlp.gate_proj",
363
+ "model.layers.34.mlp.shared_mlp.up_proj",
364
+ "model.layers.34.mlp.switch_mlp.down_proj",
365
+ "model.layers.34.mlp.switch_mlp.gate_proj",
366
+ "model.layers.34.mlp.switch_mlp.up_proj",
367
+ "model.layers.34.self_attn.k_proj",
368
+ "model.layers.34.self_attn.o_proj",
369
+ "model.layers.34.self_attn.q_proj",
370
+ "model.layers.34.self_attn.v_proj",
371
+ "model.layers.35.mlp.shared_mlp.down_proj",
372
+ "model.layers.35.mlp.shared_mlp.gate_proj",
373
+ "model.layers.35.mlp.shared_mlp.up_proj",
374
+ "model.layers.35.mlp.switch_mlp.down_proj",
375
+ "model.layers.35.mlp.switch_mlp.gate_proj",
376
+ "model.layers.35.mlp.switch_mlp.up_proj",
377
+ "model.layers.35.self_attn.k_proj",
378
+ "model.layers.35.self_attn.o_proj",
379
+ "model.layers.35.self_attn.q_proj",
380
+ "model.layers.35.self_attn.v_proj",
381
+ "model.layers.36.mlp.shared_mlp.down_proj",
382
+ "model.layers.36.mlp.shared_mlp.gate_proj",
383
+ "model.layers.36.mlp.shared_mlp.up_proj",
384
+ "model.layers.36.mlp.switch_mlp.down_proj",
385
+ "model.layers.36.mlp.switch_mlp.gate_proj",
386
+ "model.layers.36.mlp.switch_mlp.up_proj",
387
+ "model.layers.36.self_attn.k_proj",
388
+ "model.layers.36.self_attn.o_proj",
389
+ "model.layers.36.self_attn.q_proj",
390
+ "model.layers.36.self_attn.v_proj",
391
+ "model.layers.37.mlp.shared_mlp.down_proj",
392
+ "model.layers.37.mlp.shared_mlp.gate_proj",
393
+ "model.layers.37.mlp.shared_mlp.up_proj",
394
+ "model.layers.37.mlp.switch_mlp.down_proj",
395
+ "model.layers.37.mlp.switch_mlp.gate_proj",
396
+ "model.layers.37.mlp.switch_mlp.up_proj",
397
+ "model.layers.37.self_attn.k_proj",
398
+ "model.layers.37.self_attn.o_proj",
399
+ "model.layers.37.self_attn.q_proj",
400
+ "model.layers.37.self_attn.v_proj",
401
+ "model.layers.38.mlp.shared_mlp.down_proj",
402
+ "model.layers.38.mlp.shared_mlp.gate_proj",
403
+ "model.layers.38.mlp.shared_mlp.up_proj",
404
+ "model.layers.38.mlp.switch_mlp.down_proj",
405
+ "model.layers.38.mlp.switch_mlp.gate_proj",
406
+ "model.layers.38.mlp.switch_mlp.up_proj",
407
+ "model.layers.38.self_attn.k_proj",
408
+ "model.layers.38.self_attn.o_proj",
409
+ "model.layers.38.self_attn.q_proj",
410
+ "model.layers.38.self_attn.v_proj",
411
+ "model.layers.39.mlp.shared_mlp.down_proj",
412
+ "model.layers.39.mlp.shared_mlp.gate_proj",
413
+ "model.layers.39.mlp.shared_mlp.up_proj",
414
+ "model.layers.39.mlp.switch_mlp.down_proj",
415
+ "model.layers.39.mlp.switch_mlp.gate_proj",
416
+ "model.layers.39.mlp.switch_mlp.up_proj",
417
+ "model.layers.39.self_attn.k_proj",
418
+ "model.layers.39.self_attn.o_proj",
419
+ "model.layers.39.self_attn.q_proj",
420
+ "model.layers.39.self_attn.v_proj",
421
+ "model.layers.4.mlp.shared_mlp.down_proj",
422
+ "model.layers.4.mlp.shared_mlp.gate_proj",
423
+ "model.layers.4.mlp.shared_mlp.up_proj",
424
+ "model.layers.4.mlp.switch_mlp.down_proj",
425
+ "model.layers.4.mlp.switch_mlp.gate_proj",
426
+ "model.layers.4.mlp.switch_mlp.up_proj",
427
+ "model.layers.4.self_attn.k_proj",
428
+ "model.layers.4.self_attn.o_proj",
429
+ "model.layers.4.self_attn.q_proj",
430
+ "model.layers.4.self_attn.v_proj",
431
+ "model.layers.40.mlp.shared_mlp.down_proj",
432
+ "model.layers.40.mlp.shared_mlp.gate_proj",
433
+ "model.layers.40.mlp.shared_mlp.up_proj",
434
+ "model.layers.40.mlp.switch_mlp.down_proj",
435
+ "model.layers.40.mlp.switch_mlp.gate_proj",
436
+ "model.layers.40.mlp.switch_mlp.up_proj",
437
+ "model.layers.40.self_attn.k_proj",
438
+ "model.layers.40.self_attn.o_proj",
439
+ "model.layers.40.self_attn.q_proj",
440
+ "model.layers.40.self_attn.v_proj",
441
+ "model.layers.41.mlp.shared_mlp.down_proj",
442
+ "model.layers.41.mlp.shared_mlp.gate_proj",
443
+ "model.layers.41.mlp.shared_mlp.up_proj",
444
+ "model.layers.41.mlp.switch_mlp.down_proj",
445
+ "model.layers.41.mlp.switch_mlp.gate_proj",
446
+ "model.layers.41.mlp.switch_mlp.up_proj",
447
+ "model.layers.41.self_attn.k_proj",
448
+ "model.layers.41.self_attn.o_proj",
449
+ "model.layers.41.self_attn.q_proj",
450
+ "model.layers.41.self_attn.v_proj",
451
+ "model.layers.42.mlp.shared_mlp.down_proj",
452
+ "model.layers.42.mlp.shared_mlp.gate_proj",
453
+ "model.layers.42.mlp.shared_mlp.up_proj",
454
+ "model.layers.42.mlp.switch_mlp.down_proj",
455
+ "model.layers.42.mlp.switch_mlp.gate_proj",
456
+ "model.layers.42.mlp.switch_mlp.up_proj",
457
+ "model.layers.42.self_attn.k_proj",
458
+ "model.layers.42.self_attn.o_proj",
459
+ "model.layers.42.self_attn.q_proj",
460
+ "model.layers.42.self_attn.v_proj",
461
+ "model.layers.43.mlp.shared_mlp.down_proj",
462
+ "model.layers.43.mlp.shared_mlp.gate_proj",
463
+ "model.layers.43.mlp.shared_mlp.up_proj",
464
+ "model.layers.43.mlp.switch_mlp.down_proj",
465
+ "model.layers.43.mlp.switch_mlp.gate_proj",
466
+ "model.layers.43.mlp.switch_mlp.up_proj",
467
+ "model.layers.43.self_attn.k_proj",
468
+ "model.layers.43.self_attn.o_proj",
469
+ "model.layers.43.self_attn.q_proj",
470
+ "model.layers.43.self_attn.v_proj",
471
+ "model.layers.44.mlp.shared_mlp.down_proj",
472
+ "model.layers.44.mlp.shared_mlp.gate_proj",
473
+ "model.layers.44.mlp.shared_mlp.up_proj",
474
+ "model.layers.44.mlp.switch_mlp.down_proj",
475
+ "model.layers.44.mlp.switch_mlp.gate_proj",
476
+ "model.layers.44.mlp.switch_mlp.up_proj",
477
+ "model.layers.44.self_attn.k_proj",
478
+ "model.layers.44.self_attn.o_proj",
479
+ "model.layers.44.self_attn.q_proj",
480
+ "model.layers.44.self_attn.v_proj",
481
+ "model.layers.45.mlp.shared_mlp.down_proj",
482
+ "model.layers.45.mlp.shared_mlp.gate_proj",
483
+ "model.layers.45.mlp.shared_mlp.up_proj",
484
+ "model.layers.45.mlp.switch_mlp.down_proj",
485
+ "model.layers.45.mlp.switch_mlp.gate_proj",
486
+ "model.layers.45.mlp.switch_mlp.up_proj",
487
+ "model.layers.45.self_attn.k_proj",
488
+ "model.layers.45.self_attn.o_proj",
489
+ "model.layers.45.self_attn.q_proj",
490
+ "model.layers.45.self_attn.v_proj",
491
+ "model.layers.46.mlp.shared_mlp.down_proj",
492
+ "model.layers.46.mlp.shared_mlp.gate_proj",
493
+ "model.layers.46.mlp.shared_mlp.up_proj",
494
+ "model.layers.46.mlp.switch_mlp.down_proj",
495
+ "model.layers.46.mlp.switch_mlp.gate_proj",
496
+ "model.layers.46.mlp.switch_mlp.up_proj",
497
+ "model.layers.46.self_attn.k_proj",
498
+ "model.layers.46.self_attn.o_proj",
499
+ "model.layers.46.self_attn.q_proj",
500
+ "model.layers.46.self_attn.v_proj",
501
+ "model.layers.47.mlp.shared_mlp.down_proj",
502
+ "model.layers.47.mlp.shared_mlp.gate_proj",
503
+ "model.layers.47.mlp.shared_mlp.up_proj",
504
+ "model.layers.47.mlp.switch_mlp.down_proj",
505
+ "model.layers.47.mlp.switch_mlp.gate_proj",
506
+ "model.layers.47.mlp.switch_mlp.up_proj",
507
+ "model.layers.47.self_attn.k_proj",
508
+ "model.layers.47.self_attn.o_proj",
509
+ "model.layers.47.self_attn.q_proj",
510
+ "model.layers.47.self_attn.v_proj",
511
+ "model.layers.48.mlp.shared_mlp.down_proj",
512
+ "model.layers.48.mlp.shared_mlp.gate_proj",
513
+ "model.layers.48.mlp.shared_mlp.up_proj",
514
+ "model.layers.48.mlp.switch_mlp.down_proj",
515
+ "model.layers.48.mlp.switch_mlp.gate_proj",
516
+ "model.layers.48.mlp.switch_mlp.up_proj",
517
+ "model.layers.48.self_attn.k_proj",
518
+ "model.layers.48.self_attn.o_proj",
519
+ "model.layers.48.self_attn.q_proj",
520
+ "model.layers.48.self_attn.v_proj",
521
+ "model.layers.49.mlp.shared_mlp.down_proj",
522
+ "model.layers.49.mlp.shared_mlp.gate_proj",
523
+ "model.layers.49.mlp.shared_mlp.up_proj",
524
+ "model.layers.49.mlp.switch_mlp.down_proj",
525
+ "model.layers.49.mlp.switch_mlp.gate_proj",
526
+ "model.layers.49.mlp.switch_mlp.up_proj",
527
+ "model.layers.49.self_attn.k_proj",
528
+ "model.layers.49.self_attn.o_proj",
529
+ "model.layers.49.self_attn.q_proj",
530
+ "model.layers.49.self_attn.v_proj",
531
+ "model.layers.5.mlp.shared_mlp.down_proj",
532
+ "model.layers.5.mlp.shared_mlp.gate_proj",
533
+ "model.layers.5.mlp.shared_mlp.up_proj",
534
+ "model.layers.5.mlp.switch_mlp.down_proj",
535
+ "model.layers.5.mlp.switch_mlp.gate_proj",
536
+ "model.layers.5.mlp.switch_mlp.up_proj",
537
+ "model.layers.5.self_attn.k_proj",
538
+ "model.layers.5.self_attn.o_proj",
539
+ "model.layers.5.self_attn.q_proj",
540
+ "model.layers.5.self_attn.v_proj",
541
+ "model.layers.50.mlp.shared_mlp.down_proj",
542
+ "model.layers.50.mlp.shared_mlp.gate_proj",
543
+ "model.layers.50.mlp.shared_mlp.up_proj",
544
+ "model.layers.50.mlp.switch_mlp.down_proj",
545
+ "model.layers.50.mlp.switch_mlp.gate_proj",
546
+ "model.layers.50.mlp.switch_mlp.up_proj",
547
+ "model.layers.50.self_attn.k_proj",
548
+ "model.layers.50.self_attn.o_proj",
549
+ "model.layers.50.self_attn.q_proj",
550
+ "model.layers.50.self_attn.v_proj",
551
+ "model.layers.51.mlp.shared_mlp.down_proj",
552
+ "model.layers.51.mlp.shared_mlp.gate_proj",
553
+ "model.layers.51.mlp.shared_mlp.up_proj",
554
+ "model.layers.51.mlp.switch_mlp.down_proj",
555
+ "model.layers.51.mlp.switch_mlp.gate_proj",
556
+ "model.layers.51.mlp.switch_mlp.up_proj",
557
+ "model.layers.51.self_attn.k_proj",
558
+ "model.layers.51.self_attn.o_proj",
559
+ "model.layers.51.self_attn.q_proj",
560
+ "model.layers.51.self_attn.v_proj",
561
+ "model.layers.52.mlp.shared_mlp.down_proj",
562
+ "model.layers.52.mlp.shared_mlp.gate_proj",
563
+ "model.layers.52.mlp.shared_mlp.up_proj",
564
+ "model.layers.52.mlp.switch_mlp.down_proj",
565
+ "model.layers.52.mlp.switch_mlp.gate_proj",
566
+ "model.layers.52.mlp.switch_mlp.up_proj",
567
+ "model.layers.52.self_attn.k_proj",
568
+ "model.layers.52.self_attn.o_proj",
569
+ "model.layers.52.self_attn.q_proj",
570
+ "model.layers.52.self_attn.v_proj",
571
+ "model.layers.53.mlp.shared_mlp.down_proj",
572
+ "model.layers.53.mlp.shared_mlp.gate_proj",
573
+ "model.layers.53.mlp.shared_mlp.up_proj",
574
+ "model.layers.53.mlp.switch_mlp.down_proj",
575
+ "model.layers.53.mlp.switch_mlp.gate_proj",
576
+ "model.layers.53.mlp.switch_mlp.up_proj",
577
+ "model.layers.53.self_attn.k_proj",
578
+ "model.layers.53.self_attn.o_proj",
579
+ "model.layers.53.self_attn.q_proj",
580
+ "model.layers.53.self_attn.v_proj",
581
+ "model.layers.54.mlp.shared_mlp.down_proj",
582
+ "model.layers.54.mlp.shared_mlp.gate_proj",
583
+ "model.layers.54.mlp.shared_mlp.up_proj",
584
+ "model.layers.54.mlp.switch_mlp.down_proj",
585
+ "model.layers.54.mlp.switch_mlp.gate_proj",
586
+ "model.layers.54.mlp.switch_mlp.up_proj",
587
+ "model.layers.54.self_attn.k_proj",
588
+ "model.layers.54.self_attn.o_proj",
589
+ "model.layers.54.self_attn.q_proj",
590
+ "model.layers.54.self_attn.v_proj",
591
+ "model.layers.55.mlp.shared_mlp.down_proj",
592
+ "model.layers.55.mlp.shared_mlp.gate_proj",
593
+ "model.layers.55.mlp.shared_mlp.up_proj",
594
+ "model.layers.55.mlp.switch_mlp.down_proj",
595
+ "model.layers.55.mlp.switch_mlp.gate_proj",
596
+ "model.layers.55.mlp.switch_mlp.up_proj",
597
+ "model.layers.55.self_attn.k_proj",
598
+ "model.layers.55.self_attn.o_proj",
599
+ "model.layers.55.self_attn.q_proj",
600
+ "model.layers.55.self_attn.v_proj",
601
+ "model.layers.56.mlp.shared_mlp.down_proj",
602
+ "model.layers.56.mlp.shared_mlp.gate_proj",
603
+ "model.layers.56.mlp.shared_mlp.up_proj",
604
+ "model.layers.56.mlp.switch_mlp.down_proj",
605
+ "model.layers.56.mlp.switch_mlp.gate_proj",
606
+ "model.layers.56.mlp.switch_mlp.up_proj",
607
+ "model.layers.56.self_attn.k_proj",
608
+ "model.layers.56.self_attn.o_proj",
609
+ "model.layers.56.self_attn.q_proj",
610
+ "model.layers.56.self_attn.v_proj",
611
+ "model.layers.57.mlp.shared_mlp.down_proj",
612
+ "model.layers.57.mlp.shared_mlp.gate_proj",
613
+ "model.layers.57.mlp.shared_mlp.up_proj",
614
+ "model.layers.57.mlp.switch_mlp.down_proj",
615
+ "model.layers.57.mlp.switch_mlp.gate_proj",
616
+ "model.layers.57.mlp.switch_mlp.up_proj",
617
+ "model.layers.57.self_attn.k_proj",
618
+ "model.layers.57.self_attn.o_proj",
619
+ "model.layers.57.self_attn.q_proj",
620
+ "model.layers.57.self_attn.v_proj",
621
+ "model.layers.58.mlp.shared_mlp.down_proj",
622
+ "model.layers.58.mlp.shared_mlp.gate_proj",
623
+ "model.layers.58.mlp.shared_mlp.up_proj",
624
+ "model.layers.58.mlp.switch_mlp.down_proj",
625
+ "model.layers.58.mlp.switch_mlp.gate_proj",
626
+ "model.layers.58.mlp.switch_mlp.up_proj",
627
+ "model.layers.58.self_attn.k_proj",
628
+ "model.layers.58.self_attn.o_proj",
629
+ "model.layers.58.self_attn.q_proj",
630
+ "model.layers.58.self_attn.v_proj",
631
+ "model.layers.59.mlp.shared_mlp.down_proj",
632
+ "model.layers.59.mlp.shared_mlp.gate_proj",
633
+ "model.layers.59.mlp.shared_mlp.up_proj",
634
+ "model.layers.59.mlp.switch_mlp.down_proj",
635
+ "model.layers.59.mlp.switch_mlp.gate_proj",
636
+ "model.layers.59.mlp.switch_mlp.up_proj",
637
+ "model.layers.59.self_attn.k_proj",
638
+ "model.layers.59.self_attn.o_proj",
639
+ "model.layers.59.self_attn.q_proj",
640
+ "model.layers.59.self_attn.v_proj",
641
+ "model.layers.6.mlp.shared_mlp.down_proj",
642
+ "model.layers.6.mlp.shared_mlp.gate_proj",
643
+ "model.layers.6.mlp.shared_mlp.up_proj",
644
+ "model.layers.6.mlp.switch_mlp.down_proj",
645
+ "model.layers.6.mlp.switch_mlp.gate_proj",
646
+ "model.layers.6.mlp.switch_mlp.up_proj",
647
+ "model.layers.6.self_attn.k_proj",
648
+ "model.layers.6.self_attn.o_proj",
649
+ "model.layers.6.self_attn.q_proj",
650
+ "model.layers.6.self_attn.v_proj",
651
+ "model.layers.60.mlp.shared_mlp.down_proj",
652
+ "model.layers.60.mlp.shared_mlp.gate_proj",
653
+ "model.layers.60.mlp.shared_mlp.up_proj",
654
+ "model.layers.60.mlp.switch_mlp.down_proj",
655
+ "model.layers.60.mlp.switch_mlp.gate_proj",
656
+ "model.layers.60.mlp.switch_mlp.up_proj",
657
+ "model.layers.60.self_attn.k_proj",
658
+ "model.layers.60.self_attn.o_proj",
659
+ "model.layers.60.self_attn.q_proj",
660
+ "model.layers.60.self_attn.v_proj",
661
+ "model.layers.61.mlp.shared_mlp.down_proj",
662
+ "model.layers.61.mlp.shared_mlp.gate_proj",
663
+ "model.layers.61.mlp.shared_mlp.up_proj",
664
+ "model.layers.61.mlp.switch_mlp.down_proj",
665
+ "model.layers.61.mlp.switch_mlp.gate_proj",
666
+ "model.layers.61.mlp.switch_mlp.up_proj",
667
+ "model.layers.61.self_attn.k_proj",
668
+ "model.layers.61.self_attn.o_proj",
669
+ "model.layers.61.self_attn.q_proj",
670
+ "model.layers.61.self_attn.v_proj",
671
+ "model.layers.62.mlp.shared_mlp.down_proj",
672
+ "model.layers.62.mlp.shared_mlp.gate_proj",
673
+ "model.layers.62.mlp.shared_mlp.up_proj",
674
+ "model.layers.62.mlp.switch_mlp.down_proj",
675
+ "model.layers.62.mlp.switch_mlp.gate_proj",
676
+ "model.layers.62.mlp.switch_mlp.up_proj",
677
+ "model.layers.62.self_attn.k_proj",
678
+ "model.layers.62.self_attn.o_proj",
679
+ "model.layers.62.self_attn.q_proj",
680
+ "model.layers.62.self_attn.v_proj",
681
+ "model.layers.63.mlp.shared_mlp.down_proj",
682
+ "model.layers.63.mlp.shared_mlp.gate_proj",
683
+ "model.layers.63.mlp.shared_mlp.up_proj",
684
+ "model.layers.63.mlp.switch_mlp.down_proj",
685
+ "model.layers.63.mlp.switch_mlp.gate_proj",
686
+ "model.layers.63.mlp.switch_mlp.up_proj",
687
+ "model.layers.63.self_attn.k_proj",
688
+ "model.layers.63.self_attn.o_proj",
689
+ "model.layers.63.self_attn.q_proj",
690
+ "model.layers.63.self_attn.v_proj",
691
+ "model.layers.64.mlp.shared_mlp.down_proj",
692
+ "model.layers.64.mlp.shared_mlp.gate_proj",
693
+ "model.layers.64.mlp.shared_mlp.up_proj",
694
+ "model.layers.64.mlp.switch_mlp.down_proj",
695
+ "model.layers.64.mlp.switch_mlp.gate_proj",
696
+ "model.layers.64.mlp.switch_mlp.up_proj",
697
+ "model.layers.64.self_attn.k_proj",
698
+ "model.layers.64.self_attn.o_proj",
699
+ "model.layers.64.self_attn.q_proj",
700
+ "model.layers.64.self_attn.v_proj",
701
+ "model.layers.65.mlp.shared_mlp.down_proj",
702
+ "model.layers.65.mlp.shared_mlp.gate_proj",
703
+ "model.layers.65.mlp.shared_mlp.up_proj",
704
+ "model.layers.65.mlp.switch_mlp.down_proj",
705
+ "model.layers.65.mlp.switch_mlp.gate_proj",
706
+ "model.layers.65.mlp.switch_mlp.up_proj",
707
+ "model.layers.65.self_attn.k_proj",
708
+ "model.layers.65.self_attn.o_proj",
709
+ "model.layers.65.self_attn.q_proj",
710
+ "model.layers.65.self_attn.v_proj",
711
+ "model.layers.66.mlp.shared_mlp.down_proj",
712
+ "model.layers.66.mlp.shared_mlp.gate_proj",
713
+ "model.layers.66.mlp.shared_mlp.up_proj",
714
+ "model.layers.66.mlp.switch_mlp.down_proj",
715
+ "model.layers.66.mlp.switch_mlp.gate_proj",
716
+ "model.layers.66.mlp.switch_mlp.up_proj",
717
+ "model.layers.66.self_attn.k_proj",
718
+ "model.layers.66.self_attn.o_proj",
719
+ "model.layers.66.self_attn.q_proj",
720
+ "model.layers.66.self_attn.v_proj",
721
+ "model.layers.67.mlp.shared_mlp.down_proj",
722
+ "model.layers.67.mlp.shared_mlp.gate_proj",
723
+ "model.layers.67.mlp.shared_mlp.up_proj",
724
+ "model.layers.67.mlp.switch_mlp.down_proj",
725
+ "model.layers.67.mlp.switch_mlp.gate_proj",
726
+ "model.layers.67.mlp.switch_mlp.up_proj",
727
+ "model.layers.67.self_attn.k_proj",
728
+ "model.layers.67.self_attn.o_proj",
729
+ "model.layers.67.self_attn.q_proj",
730
+ "model.layers.67.self_attn.v_proj",
731
+ "model.layers.68.mlp.shared_mlp.down_proj",
732
+ "model.layers.68.mlp.shared_mlp.gate_proj",
733
+ "model.layers.68.mlp.shared_mlp.up_proj",
734
+ "model.layers.68.mlp.switch_mlp.down_proj",
735
+ "model.layers.68.mlp.switch_mlp.gate_proj",
736
+ "model.layers.68.mlp.switch_mlp.up_proj",
737
+ "model.layers.68.self_attn.k_proj",
738
+ "model.layers.68.self_attn.o_proj",
739
+ "model.layers.68.self_attn.q_proj",
740
+ "model.layers.68.self_attn.v_proj",
741
+ "model.layers.69.mlp.shared_mlp.down_proj",
742
+ "model.layers.69.mlp.shared_mlp.gate_proj",
743
+ "model.layers.69.mlp.shared_mlp.up_proj",
744
+ "model.layers.69.mlp.switch_mlp.down_proj",
745
+ "model.layers.69.mlp.switch_mlp.gate_proj",
746
+ "model.layers.69.mlp.switch_mlp.up_proj",
747
+ "model.layers.69.self_attn.k_proj",
748
+ "model.layers.69.self_attn.o_proj",
749
+ "model.layers.69.self_attn.q_proj",
750
+ "model.layers.69.self_attn.v_proj",
751
+ "model.layers.7.mlp.shared_mlp.down_proj",
752
+ "model.layers.7.mlp.shared_mlp.gate_proj",
753
+ "model.layers.7.mlp.shared_mlp.up_proj",
754
+ "model.layers.7.mlp.switch_mlp.down_proj",
755
+ "model.layers.7.mlp.switch_mlp.gate_proj",
756
+ "model.layers.7.mlp.switch_mlp.up_proj",
757
+ "model.layers.7.self_attn.k_proj",
758
+ "model.layers.7.self_attn.o_proj",
759
+ "model.layers.7.self_attn.q_proj",
760
+ "model.layers.7.self_attn.v_proj",
761
+ "model.layers.70.mlp.shared_mlp.down_proj",
762
+ "model.layers.70.mlp.shared_mlp.gate_proj",
763
+ "model.layers.70.mlp.shared_mlp.up_proj",
764
+ "model.layers.70.mlp.switch_mlp.down_proj",
765
+ "model.layers.70.mlp.switch_mlp.gate_proj",
766
+ "model.layers.70.mlp.switch_mlp.up_proj",
767
+ "model.layers.70.self_attn.k_proj",
768
+ "model.layers.70.self_attn.o_proj",
769
+ "model.layers.70.self_attn.q_proj",
770
+ "model.layers.70.self_attn.v_proj",
771
+ "model.layers.71.mlp.shared_mlp.down_proj",
772
+ "model.layers.71.mlp.shared_mlp.gate_proj",
773
+ "model.layers.71.mlp.shared_mlp.up_proj",
774
+ "model.layers.71.mlp.switch_mlp.down_proj",
775
+ "model.layers.71.mlp.switch_mlp.gate_proj",
776
+ "model.layers.71.mlp.switch_mlp.up_proj",
777
+ "model.layers.71.self_attn.k_proj",
778
+ "model.layers.71.self_attn.o_proj",
779
+ "model.layers.71.self_attn.q_proj",
780
+ "model.layers.71.self_attn.v_proj",
781
+ "model.layers.72.mlp.shared_mlp.down_proj",
782
+ "model.layers.72.mlp.shared_mlp.gate_proj",
783
+ "model.layers.72.mlp.shared_mlp.up_proj",
784
+ "model.layers.72.mlp.switch_mlp.down_proj",
785
+ "model.layers.72.mlp.switch_mlp.gate_proj",
786
+ "model.layers.72.mlp.switch_mlp.up_proj",
787
+ "model.layers.72.self_attn.k_proj",
788
+ "model.layers.72.self_attn.o_proj",
789
+ "model.layers.72.self_attn.q_proj",
790
+ "model.layers.72.self_attn.v_proj",
791
+ "model.layers.73.mlp.shared_mlp.down_proj",
792
+ "model.layers.73.mlp.shared_mlp.gate_proj",
793
+ "model.layers.73.mlp.shared_mlp.up_proj",
794
+ "model.layers.73.mlp.switch_mlp.down_proj",
795
+ "model.layers.73.mlp.switch_mlp.gate_proj",
796
+ "model.layers.73.mlp.switch_mlp.up_proj",
797
+ "model.layers.73.self_attn.k_proj",
798
+ "model.layers.73.self_attn.o_proj",
799
+ "model.layers.73.self_attn.q_proj",
800
+ "model.layers.73.self_attn.v_proj",
801
+ "model.layers.74.mlp.shared_mlp.down_proj",
802
+ "model.layers.74.mlp.shared_mlp.gate_proj",
803
+ "model.layers.74.mlp.shared_mlp.up_proj",
804
+ "model.layers.74.mlp.switch_mlp.down_proj",
805
+ "model.layers.74.mlp.switch_mlp.gate_proj",
806
+ "model.layers.74.mlp.switch_mlp.up_proj",
807
+ "model.layers.74.self_attn.k_proj",
808
+ "model.layers.74.self_attn.o_proj",
809
+ "model.layers.74.self_attn.q_proj",
810
+ "model.layers.74.self_attn.v_proj",
811
+ "model.layers.75.mlp.shared_mlp.down_proj",
812
+ "model.layers.75.mlp.shared_mlp.gate_proj",
813
+ "model.layers.75.mlp.shared_mlp.up_proj",
814
+ "model.layers.75.mlp.switch_mlp.down_proj",
815
+ "model.layers.75.mlp.switch_mlp.gate_proj",
816
+ "model.layers.75.mlp.switch_mlp.up_proj",
817
+ "model.layers.75.self_attn.k_proj",
818
+ "model.layers.75.self_attn.o_proj",
819
+ "model.layers.75.self_attn.q_proj",
820
+ "model.layers.75.self_attn.v_proj",
821
+ "model.layers.76.mlp.shared_mlp.down_proj",
822
+ "model.layers.76.mlp.shared_mlp.gate_proj",
823
+ "model.layers.76.mlp.shared_mlp.up_proj",
824
+ "model.layers.76.mlp.switch_mlp.down_proj",
825
+ "model.layers.76.mlp.switch_mlp.gate_proj",
826
+ "model.layers.76.mlp.switch_mlp.up_proj",
827
+ "model.layers.76.self_attn.k_proj",
828
+ "model.layers.76.self_attn.o_proj",
829
+ "model.layers.76.self_attn.q_proj",
830
+ "model.layers.76.self_attn.v_proj",
831
+ "model.layers.77.mlp.shared_mlp.down_proj",
832
+ "model.layers.77.mlp.shared_mlp.gate_proj",
833
+ "model.layers.77.mlp.shared_mlp.up_proj",
834
+ "model.layers.77.mlp.switch_mlp.down_proj",
835
+ "model.layers.77.mlp.switch_mlp.gate_proj",
836
+ "model.layers.77.mlp.switch_mlp.up_proj",
837
+ "model.layers.77.self_attn.k_proj",
838
+ "model.layers.77.self_attn.o_proj",
839
+ "model.layers.77.self_attn.q_proj",
840
+ "model.layers.77.self_attn.v_proj",
841
+ "model.layers.78.mlp.shared_mlp.down_proj",
842
+ "model.layers.78.mlp.shared_mlp.gate_proj",
843
+ "model.layers.78.mlp.shared_mlp.up_proj",
844
+ "model.layers.78.mlp.switch_mlp.down_proj",
845
+ "model.layers.78.mlp.switch_mlp.gate_proj",
846
+ "model.layers.78.mlp.switch_mlp.up_proj",
847
+ "model.layers.78.self_attn.k_proj",
848
+ "model.layers.78.self_attn.o_proj",
849
+ "model.layers.78.self_attn.q_proj",
850
+ "model.layers.78.self_attn.v_proj",
851
+ "model.layers.79.mlp.shared_mlp.down_proj",
852
+ "model.layers.79.mlp.shared_mlp.gate_proj",
853
+ "model.layers.79.mlp.shared_mlp.up_proj",
854
+ "model.layers.79.mlp.switch_mlp.down_proj",
855
+ "model.layers.79.mlp.switch_mlp.gate_proj",
856
+ "model.layers.79.mlp.switch_mlp.up_proj",
857
+ "model.layers.79.self_attn.k_proj",
858
+ "model.layers.79.self_attn.o_proj",
859
+ "model.layers.79.self_attn.q_proj",
860
+ "model.layers.79.self_attn.v_proj",
861
+ "model.layers.8.mlp.shared_mlp.down_proj",
862
+ "model.layers.8.mlp.shared_mlp.gate_proj",
863
+ "model.layers.8.mlp.shared_mlp.up_proj",
864
+ "model.layers.8.mlp.switch_mlp.down_proj",
865
+ "model.layers.8.mlp.switch_mlp.gate_proj",
866
+ "model.layers.8.mlp.switch_mlp.up_proj",
867
+ "model.layers.8.self_attn.k_proj",
868
+ "model.layers.8.self_attn.o_proj",
869
+ "model.layers.8.self_attn.q_proj",
870
+ "model.layers.8.self_attn.v_proj",
871
+ "model.layers.9.mlp.shared_mlp.down_proj",
872
+ "model.layers.9.mlp.shared_mlp.gate_proj",
873
+ "model.layers.9.mlp.shared_mlp.up_proj",
874
+ "model.layers.9.mlp.switch_mlp.down_proj",
875
+ "model.layers.9.mlp.switch_mlp.gate_proj",
876
+ "model.layers.9.mlp.switch_mlp.up_proj",
877
+ "model.layers.9.self_attn.k_proj",
878
+ "model.layers.9.self_attn.o_proj",
879
+ "model.layers.9.self_attn.q_proj",
880
+ "model.layers.9.self_attn.v_proj"
881
+ ],
882
+ "missing": [
883
+ "lm_head",
884
+ "model.embed_tokens"
885
+ ],
886
+ "mismatched": [],
887
+ "zero_count_experts": 0
888
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
The diff for this file is too large to render. See raw diff