twodgirl commited on
Commit
25924b4
1 Parent(s): b26054a

Upload flux dev transformer.

Browse files
flux-fp8/config.json ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "FluxTransformer2DModel",
3
+ "_diffusers_version": "0.30.0.dev0",
4
+ "_name_or_path": "./flux/transformer",
5
+ "attention_head_dim": 128,
6
+ "axes_dims_rope": [
7
+ 16,
8
+ 56,
9
+ 56
10
+ ],
11
+ "guidance_embeds": true,
12
+ "in_channels": 64,
13
+ "joint_attention_dim": 4096,
14
+ "num_attention_heads": 24,
15
+ "num_layers": 19,
16
+ "num_single_layers": 38,
17
+ "patch_size": 1,
18
+ "pooled_projection_dim": 768
19
+ }
flux-fp8/diffusion_pytorch_model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c4bdb0288237a538ab0ae2bda770e81a93843f42562065afc818b8addfaad685
3
+ size 9974223248
flux-fp8/diffusion_pytorch_model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:84e4dcf2777246ff69a2bafd5599b943005c1cc79c6385a3dff700f355e06c32
3
+ size 1936709200
flux-fp8/diffusion_pytorch_model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
flux-fp8/quanto_qmap.json ADDED
@@ -0,0 +1,2018 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "time_text_embed.timestep_embedder.linear_1": {
3
+ "weights": "qfloat8_e4m3fn",
4
+ "activations": "none"
5
+ },
6
+ "time_text_embed.timestep_embedder.linear_2": {
7
+ "weights": "qfloat8_e4m3fn",
8
+ "activations": "none"
9
+ },
10
+ "time_text_embed.guidance_embedder.linear_1": {
11
+ "weights": "qfloat8_e4m3fn",
12
+ "activations": "none"
13
+ },
14
+ "time_text_embed.guidance_embedder.linear_2": {
15
+ "weights": "qfloat8_e4m3fn",
16
+ "activations": "none"
17
+ },
18
+ "time_text_embed.text_embedder.linear_1": {
19
+ "weights": "qfloat8_e4m3fn",
20
+ "activations": "none"
21
+ },
22
+ "time_text_embed.text_embedder.linear_2": {
23
+ "weights": "qfloat8_e4m3fn",
24
+ "activations": "none"
25
+ },
26
+ "context_embedder": {
27
+ "weights": "qfloat8_e4m3fn",
28
+ "activations": "none"
29
+ },
30
+ "x_embedder": {
31
+ "weights": "qfloat8_e4m3fn",
32
+ "activations": "none"
33
+ },
34
+ "transformer_blocks.0.norm1.linear": {
35
+ "weights": "qfloat8_e4m3fn",
36
+ "activations": "none"
37
+ },
38
+ "transformer_blocks.0.norm1_context.linear": {
39
+ "weights": "qfloat8_e4m3fn",
40
+ "activations": "none"
41
+ },
42
+ "transformer_blocks.0.attn.to_q": {
43
+ "weights": "qfloat8_e4m3fn",
44
+ "activations": "none"
45
+ },
46
+ "transformer_blocks.0.attn.to_k": {
47
+ "weights": "qfloat8_e4m3fn",
48
+ "activations": "none"
49
+ },
50
+ "transformer_blocks.0.attn.to_v": {
51
+ "weights": "qfloat8_e4m3fn",
52
+ "activations": "none"
53
+ },
54
+ "transformer_blocks.0.attn.add_k_proj": {
55
+ "weights": "qfloat8_e4m3fn",
56
+ "activations": "none"
57
+ },
58
+ "transformer_blocks.0.attn.add_v_proj": {
59
+ "weights": "qfloat8_e4m3fn",
60
+ "activations": "none"
61
+ },
62
+ "transformer_blocks.0.attn.add_q_proj": {
63
+ "weights": "qfloat8_e4m3fn",
64
+ "activations": "none"
65
+ },
66
+ "transformer_blocks.0.attn.to_out.0": {
67
+ "weights": "qfloat8_e4m3fn",
68
+ "activations": "none"
69
+ },
70
+ "transformer_blocks.0.attn.to_add_out": {
71
+ "weights": "qfloat8_e4m3fn",
72
+ "activations": "none"
73
+ },
74
+ "transformer_blocks.0.ff.net.0.proj": {
75
+ "weights": "qfloat8_e4m3fn",
76
+ "activations": "none"
77
+ },
78
+ "transformer_blocks.0.ff.net.2": {
79
+ "weights": "qfloat8_e4m3fn",
80
+ "activations": "none"
81
+ },
82
+ "transformer_blocks.0.ff_context.net.0.proj": {
83
+ "weights": "qfloat8_e4m3fn",
84
+ "activations": "none"
85
+ },
86
+ "transformer_blocks.0.ff_context.net.2": {
87
+ "weights": "qfloat8_e4m3fn",
88
+ "activations": "none"
89
+ },
90
+ "transformer_blocks.1.norm1.linear": {
91
+ "weights": "qfloat8_e4m3fn",
92
+ "activations": "none"
93
+ },
94
+ "transformer_blocks.1.norm1_context.linear": {
95
+ "weights": "qfloat8_e4m3fn",
96
+ "activations": "none"
97
+ },
98
+ "transformer_blocks.1.attn.to_q": {
99
+ "weights": "qfloat8_e4m3fn",
100
+ "activations": "none"
101
+ },
102
+ "transformer_blocks.1.attn.to_k": {
103
+ "weights": "qfloat8_e4m3fn",
104
+ "activations": "none"
105
+ },
106
+ "transformer_blocks.1.attn.to_v": {
107
+ "weights": "qfloat8_e4m3fn",
108
+ "activations": "none"
109
+ },
110
+ "transformer_blocks.1.attn.add_k_proj": {
111
+ "weights": "qfloat8_e4m3fn",
112
+ "activations": "none"
113
+ },
114
+ "transformer_blocks.1.attn.add_v_proj": {
115
+ "weights": "qfloat8_e4m3fn",
116
+ "activations": "none"
117
+ },
118
+ "transformer_blocks.1.attn.add_q_proj": {
119
+ "weights": "qfloat8_e4m3fn",
120
+ "activations": "none"
121
+ },
122
+ "transformer_blocks.1.attn.to_out.0": {
123
+ "weights": "qfloat8_e4m3fn",
124
+ "activations": "none"
125
+ },
126
+ "transformer_blocks.1.attn.to_add_out": {
127
+ "weights": "qfloat8_e4m3fn",
128
+ "activations": "none"
129
+ },
130
+ "transformer_blocks.1.ff.net.0.proj": {
131
+ "weights": "qfloat8_e4m3fn",
132
+ "activations": "none"
133
+ },
134
+ "transformer_blocks.1.ff.net.2": {
135
+ "weights": "qfloat8_e4m3fn",
136
+ "activations": "none"
137
+ },
138
+ "transformer_blocks.1.ff_context.net.0.proj": {
139
+ "weights": "qfloat8_e4m3fn",
140
+ "activations": "none"
141
+ },
142
+ "transformer_blocks.1.ff_context.net.2": {
143
+ "weights": "qfloat8_e4m3fn",
144
+ "activations": "none"
145
+ },
146
+ "transformer_blocks.2.norm1.linear": {
147
+ "weights": "qfloat8_e4m3fn",
148
+ "activations": "none"
149
+ },
150
+ "transformer_blocks.2.norm1_context.linear": {
151
+ "weights": "qfloat8_e4m3fn",
152
+ "activations": "none"
153
+ },
154
+ "transformer_blocks.2.attn.to_q": {
155
+ "weights": "qfloat8_e4m3fn",
156
+ "activations": "none"
157
+ },
158
+ "transformer_blocks.2.attn.to_k": {
159
+ "weights": "qfloat8_e4m3fn",
160
+ "activations": "none"
161
+ },
162
+ "transformer_blocks.2.attn.to_v": {
163
+ "weights": "qfloat8_e4m3fn",
164
+ "activations": "none"
165
+ },
166
+ "transformer_blocks.2.attn.add_k_proj": {
167
+ "weights": "qfloat8_e4m3fn",
168
+ "activations": "none"
169
+ },
170
+ "transformer_blocks.2.attn.add_v_proj": {
171
+ "weights": "qfloat8_e4m3fn",
172
+ "activations": "none"
173
+ },
174
+ "transformer_blocks.2.attn.add_q_proj": {
175
+ "weights": "qfloat8_e4m3fn",
176
+ "activations": "none"
177
+ },
178
+ "transformer_blocks.2.attn.to_out.0": {
179
+ "weights": "qfloat8_e4m3fn",
180
+ "activations": "none"
181
+ },
182
+ "transformer_blocks.2.attn.to_add_out": {
183
+ "weights": "qfloat8_e4m3fn",
184
+ "activations": "none"
185
+ },
186
+ "transformer_blocks.2.ff.net.0.proj": {
187
+ "weights": "qfloat8_e4m3fn",
188
+ "activations": "none"
189
+ },
190
+ "transformer_blocks.2.ff.net.2": {
191
+ "weights": "qfloat8_e4m3fn",
192
+ "activations": "none"
193
+ },
194
+ "transformer_blocks.2.ff_context.net.0.proj": {
195
+ "weights": "qfloat8_e4m3fn",
196
+ "activations": "none"
197
+ },
198
+ "transformer_blocks.2.ff_context.net.2": {
199
+ "weights": "qfloat8_e4m3fn",
200
+ "activations": "none"
201
+ },
202
+ "transformer_blocks.3.norm1.linear": {
203
+ "weights": "qfloat8_e4m3fn",
204
+ "activations": "none"
205
+ },
206
+ "transformer_blocks.3.norm1_context.linear": {
207
+ "weights": "qfloat8_e4m3fn",
208
+ "activations": "none"
209
+ },
210
+ "transformer_blocks.3.attn.to_q": {
211
+ "weights": "qfloat8_e4m3fn",
212
+ "activations": "none"
213
+ },
214
+ "transformer_blocks.3.attn.to_k": {
215
+ "weights": "qfloat8_e4m3fn",
216
+ "activations": "none"
217
+ },
218
+ "transformer_blocks.3.attn.to_v": {
219
+ "weights": "qfloat8_e4m3fn",
220
+ "activations": "none"
221
+ },
222
+ "transformer_blocks.3.attn.add_k_proj": {
223
+ "weights": "qfloat8_e4m3fn",
224
+ "activations": "none"
225
+ },
226
+ "transformer_blocks.3.attn.add_v_proj": {
227
+ "weights": "qfloat8_e4m3fn",
228
+ "activations": "none"
229
+ },
230
+ "transformer_blocks.3.attn.add_q_proj": {
231
+ "weights": "qfloat8_e4m3fn",
232
+ "activations": "none"
233
+ },
234
+ "transformer_blocks.3.attn.to_out.0": {
235
+ "weights": "qfloat8_e4m3fn",
236
+ "activations": "none"
237
+ },
238
+ "transformer_blocks.3.attn.to_add_out": {
239
+ "weights": "qfloat8_e4m3fn",
240
+ "activations": "none"
241
+ },
242
+ "transformer_blocks.3.ff.net.0.proj": {
243
+ "weights": "qfloat8_e4m3fn",
244
+ "activations": "none"
245
+ },
246
+ "transformer_blocks.3.ff.net.2": {
247
+ "weights": "qfloat8_e4m3fn",
248
+ "activations": "none"
249
+ },
250
+ "transformer_blocks.3.ff_context.net.0.proj": {
251
+ "weights": "qfloat8_e4m3fn",
252
+ "activations": "none"
253
+ },
254
+ "transformer_blocks.3.ff_context.net.2": {
255
+ "weights": "qfloat8_e4m3fn",
256
+ "activations": "none"
257
+ },
258
+ "transformer_blocks.4.norm1.linear": {
259
+ "weights": "qfloat8_e4m3fn",
260
+ "activations": "none"
261
+ },
262
+ "transformer_blocks.4.norm1_context.linear": {
263
+ "weights": "qfloat8_e4m3fn",
264
+ "activations": "none"
265
+ },
266
+ "transformer_blocks.4.attn.to_q": {
267
+ "weights": "qfloat8_e4m3fn",
268
+ "activations": "none"
269
+ },
270
+ "transformer_blocks.4.attn.to_k": {
271
+ "weights": "qfloat8_e4m3fn",
272
+ "activations": "none"
273
+ },
274
+ "transformer_blocks.4.attn.to_v": {
275
+ "weights": "qfloat8_e4m3fn",
276
+ "activations": "none"
277
+ },
278
+ "transformer_blocks.4.attn.add_k_proj": {
279
+ "weights": "qfloat8_e4m3fn",
280
+ "activations": "none"
281
+ },
282
+ "transformer_blocks.4.attn.add_v_proj": {
283
+ "weights": "qfloat8_e4m3fn",
284
+ "activations": "none"
285
+ },
286
+ "transformer_blocks.4.attn.add_q_proj": {
287
+ "weights": "qfloat8_e4m3fn",
288
+ "activations": "none"
289
+ },
290
+ "transformer_blocks.4.attn.to_out.0": {
291
+ "weights": "qfloat8_e4m3fn",
292
+ "activations": "none"
293
+ },
294
+ "transformer_blocks.4.attn.to_add_out": {
295
+ "weights": "qfloat8_e4m3fn",
296
+ "activations": "none"
297
+ },
298
+ "transformer_blocks.4.ff.net.0.proj": {
299
+ "weights": "qfloat8_e4m3fn",
300
+ "activations": "none"
301
+ },
302
+ "transformer_blocks.4.ff.net.2": {
303
+ "weights": "qfloat8_e4m3fn",
304
+ "activations": "none"
305
+ },
306
+ "transformer_blocks.4.ff_context.net.0.proj": {
307
+ "weights": "qfloat8_e4m3fn",
308
+ "activations": "none"
309
+ },
310
+ "transformer_blocks.4.ff_context.net.2": {
311
+ "weights": "qfloat8_e4m3fn",
312
+ "activations": "none"
313
+ },
314
+ "transformer_blocks.5.norm1.linear": {
315
+ "weights": "qfloat8_e4m3fn",
316
+ "activations": "none"
317
+ },
318
+ "transformer_blocks.5.norm1_context.linear": {
319
+ "weights": "qfloat8_e4m3fn",
320
+ "activations": "none"
321
+ },
322
+ "transformer_blocks.5.attn.to_q": {
323
+ "weights": "qfloat8_e4m3fn",
324
+ "activations": "none"
325
+ },
326
+ "transformer_blocks.5.attn.to_k": {
327
+ "weights": "qfloat8_e4m3fn",
328
+ "activations": "none"
329
+ },
330
+ "transformer_blocks.5.attn.to_v": {
331
+ "weights": "qfloat8_e4m3fn",
332
+ "activations": "none"
333
+ },
334
+ "transformer_blocks.5.attn.add_k_proj": {
335
+ "weights": "qfloat8_e4m3fn",
336
+ "activations": "none"
337
+ },
338
+ "transformer_blocks.5.attn.add_v_proj": {
339
+ "weights": "qfloat8_e4m3fn",
340
+ "activations": "none"
341
+ },
342
+ "transformer_blocks.5.attn.add_q_proj": {
343
+ "weights": "qfloat8_e4m3fn",
344
+ "activations": "none"
345
+ },
346
+ "transformer_blocks.5.attn.to_out.0": {
347
+ "weights": "qfloat8_e4m3fn",
348
+ "activations": "none"
349
+ },
350
+ "transformer_blocks.5.attn.to_add_out": {
351
+ "weights": "qfloat8_e4m3fn",
352
+ "activations": "none"
353
+ },
354
+ "transformer_blocks.5.ff.net.0.proj": {
355
+ "weights": "qfloat8_e4m3fn",
356
+ "activations": "none"
357
+ },
358
+ "transformer_blocks.5.ff.net.2": {
359
+ "weights": "qfloat8_e4m3fn",
360
+ "activations": "none"
361
+ },
362
+ "transformer_blocks.5.ff_context.net.0.proj": {
363
+ "weights": "qfloat8_e4m3fn",
364
+ "activations": "none"
365
+ },
366
+ "transformer_blocks.5.ff_context.net.2": {
367
+ "weights": "qfloat8_e4m3fn",
368
+ "activations": "none"
369
+ },
370
+ "transformer_blocks.6.norm1.linear": {
371
+ "weights": "qfloat8_e4m3fn",
372
+ "activations": "none"
373
+ },
374
+ "transformer_blocks.6.norm1_context.linear": {
375
+ "weights": "qfloat8_e4m3fn",
376
+ "activations": "none"
377
+ },
378
+ "transformer_blocks.6.attn.to_q": {
379
+ "weights": "qfloat8_e4m3fn",
380
+ "activations": "none"
381
+ },
382
+ "transformer_blocks.6.attn.to_k": {
383
+ "weights": "qfloat8_e4m3fn",
384
+ "activations": "none"
385
+ },
386
+ "transformer_blocks.6.attn.to_v": {
387
+ "weights": "qfloat8_e4m3fn",
388
+ "activations": "none"
389
+ },
390
+ "transformer_blocks.6.attn.add_k_proj": {
391
+ "weights": "qfloat8_e4m3fn",
392
+ "activations": "none"
393
+ },
394
+ "transformer_blocks.6.attn.add_v_proj": {
395
+ "weights": "qfloat8_e4m3fn",
396
+ "activations": "none"
397
+ },
398
+ "transformer_blocks.6.attn.add_q_proj": {
399
+ "weights": "qfloat8_e4m3fn",
400
+ "activations": "none"
401
+ },
402
+ "transformer_blocks.6.attn.to_out.0": {
403
+ "weights": "qfloat8_e4m3fn",
404
+ "activations": "none"
405
+ },
406
+ "transformer_blocks.6.attn.to_add_out": {
407
+ "weights": "qfloat8_e4m3fn",
408
+ "activations": "none"
409
+ },
410
+ "transformer_blocks.6.ff.net.0.proj": {
411
+ "weights": "qfloat8_e4m3fn",
412
+ "activations": "none"
413
+ },
414
+ "transformer_blocks.6.ff.net.2": {
415
+ "weights": "qfloat8_e4m3fn",
416
+ "activations": "none"
417
+ },
418
+ "transformer_blocks.6.ff_context.net.0.proj": {
419
+ "weights": "qfloat8_e4m3fn",
420
+ "activations": "none"
421
+ },
422
+ "transformer_blocks.6.ff_context.net.2": {
423
+ "weights": "qfloat8_e4m3fn",
424
+ "activations": "none"
425
+ },
426
+ "transformer_blocks.7.norm1.linear": {
427
+ "weights": "qfloat8_e4m3fn",
428
+ "activations": "none"
429
+ },
430
+ "transformer_blocks.7.norm1_context.linear": {
431
+ "weights": "qfloat8_e4m3fn",
432
+ "activations": "none"
433
+ },
434
+ "transformer_blocks.7.attn.to_q": {
435
+ "weights": "qfloat8_e4m3fn",
436
+ "activations": "none"
437
+ },
438
+ "transformer_blocks.7.attn.to_k": {
439
+ "weights": "qfloat8_e4m3fn",
440
+ "activations": "none"
441
+ },
442
+ "transformer_blocks.7.attn.to_v": {
443
+ "weights": "qfloat8_e4m3fn",
444
+ "activations": "none"
445
+ },
446
+ "transformer_blocks.7.attn.add_k_proj": {
447
+ "weights": "qfloat8_e4m3fn",
448
+ "activations": "none"
449
+ },
450
+ "transformer_blocks.7.attn.add_v_proj": {
451
+ "weights": "qfloat8_e4m3fn",
452
+ "activations": "none"
453
+ },
454
+ "transformer_blocks.7.attn.add_q_proj": {
455
+ "weights": "qfloat8_e4m3fn",
456
+ "activations": "none"
457
+ },
458
+ "transformer_blocks.7.attn.to_out.0": {
459
+ "weights": "qfloat8_e4m3fn",
460
+ "activations": "none"
461
+ },
462
+ "transformer_blocks.7.attn.to_add_out": {
463
+ "weights": "qfloat8_e4m3fn",
464
+ "activations": "none"
465
+ },
466
+ "transformer_blocks.7.ff.net.0.proj": {
467
+ "weights": "qfloat8_e4m3fn",
468
+ "activations": "none"
469
+ },
470
+ "transformer_blocks.7.ff.net.2": {
471
+ "weights": "qfloat8_e4m3fn",
472
+ "activations": "none"
473
+ },
474
+ "transformer_blocks.7.ff_context.net.0.proj": {
475
+ "weights": "qfloat8_e4m3fn",
476
+ "activations": "none"
477
+ },
478
+ "transformer_blocks.7.ff_context.net.2": {
479
+ "weights": "qfloat8_e4m3fn",
480
+ "activations": "none"
481
+ },
482
+ "transformer_blocks.8.norm1.linear": {
483
+ "weights": "qfloat8_e4m3fn",
484
+ "activations": "none"
485
+ },
486
+ "transformer_blocks.8.norm1_context.linear": {
487
+ "weights": "qfloat8_e4m3fn",
488
+ "activations": "none"
489
+ },
490
+ "transformer_blocks.8.attn.to_q": {
491
+ "weights": "qfloat8_e4m3fn",
492
+ "activations": "none"
493
+ },
494
+ "transformer_blocks.8.attn.to_k": {
495
+ "weights": "qfloat8_e4m3fn",
496
+ "activations": "none"
497
+ },
498
+ "transformer_blocks.8.attn.to_v": {
499
+ "weights": "qfloat8_e4m3fn",
500
+ "activations": "none"
501
+ },
502
+ "transformer_blocks.8.attn.add_k_proj": {
503
+ "weights": "qfloat8_e4m3fn",
504
+ "activations": "none"
505
+ },
506
+ "transformer_blocks.8.attn.add_v_proj": {
507
+ "weights": "qfloat8_e4m3fn",
508
+ "activations": "none"
509
+ },
510
+ "transformer_blocks.8.attn.add_q_proj": {
511
+ "weights": "qfloat8_e4m3fn",
512
+ "activations": "none"
513
+ },
514
+ "transformer_blocks.8.attn.to_out.0": {
515
+ "weights": "qfloat8_e4m3fn",
516
+ "activations": "none"
517
+ },
518
+ "transformer_blocks.8.attn.to_add_out": {
519
+ "weights": "qfloat8_e4m3fn",
520
+ "activations": "none"
521
+ },
522
+ "transformer_blocks.8.ff.net.0.proj": {
523
+ "weights": "qfloat8_e4m3fn",
524
+ "activations": "none"
525
+ },
526
+ "transformer_blocks.8.ff.net.2": {
527
+ "weights": "qfloat8_e4m3fn",
528
+ "activations": "none"
529
+ },
530
+ "transformer_blocks.8.ff_context.net.0.proj": {
531
+ "weights": "qfloat8_e4m3fn",
532
+ "activations": "none"
533
+ },
534
+ "transformer_blocks.8.ff_context.net.2": {
535
+ "weights": "qfloat8_e4m3fn",
536
+ "activations": "none"
537
+ },
538
+ "transformer_blocks.9.norm1.linear": {
539
+ "weights": "qfloat8_e4m3fn",
540
+ "activations": "none"
541
+ },
542
+ "transformer_blocks.9.norm1_context.linear": {
543
+ "weights": "qfloat8_e4m3fn",
544
+ "activations": "none"
545
+ },
546
+ "transformer_blocks.9.attn.to_q": {
547
+ "weights": "qfloat8_e4m3fn",
548
+ "activations": "none"
549
+ },
550
+ "transformer_blocks.9.attn.to_k": {
551
+ "weights": "qfloat8_e4m3fn",
552
+ "activations": "none"
553
+ },
554
+ "transformer_blocks.9.attn.to_v": {
555
+ "weights": "qfloat8_e4m3fn",
556
+ "activations": "none"
557
+ },
558
+ "transformer_blocks.9.attn.add_k_proj": {
559
+ "weights": "qfloat8_e4m3fn",
560
+ "activations": "none"
561
+ },
562
+ "transformer_blocks.9.attn.add_v_proj": {
563
+ "weights": "qfloat8_e4m3fn",
564
+ "activations": "none"
565
+ },
566
+ "transformer_blocks.9.attn.add_q_proj": {
567
+ "weights": "qfloat8_e4m3fn",
568
+ "activations": "none"
569
+ },
570
+ "transformer_blocks.9.attn.to_out.0": {
571
+ "weights": "qfloat8_e4m3fn",
572
+ "activations": "none"
573
+ },
574
+ "transformer_blocks.9.attn.to_add_out": {
575
+ "weights": "qfloat8_e4m3fn",
576
+ "activations": "none"
577
+ },
578
+ "transformer_blocks.9.ff.net.0.proj": {
579
+ "weights": "qfloat8_e4m3fn",
580
+ "activations": "none"
581
+ },
582
+ "transformer_blocks.9.ff.net.2": {
583
+ "weights": "qfloat8_e4m3fn",
584
+ "activations": "none"
585
+ },
586
+ "transformer_blocks.9.ff_context.net.0.proj": {
587
+ "weights": "qfloat8_e4m3fn",
588
+ "activations": "none"
589
+ },
590
+ "transformer_blocks.9.ff_context.net.2": {
591
+ "weights": "qfloat8_e4m3fn",
592
+ "activations": "none"
593
+ },
594
+ "transformer_blocks.10.norm1.linear": {
595
+ "weights": "qfloat8_e4m3fn",
596
+ "activations": "none"
597
+ },
598
+ "transformer_blocks.10.norm1_context.linear": {
599
+ "weights": "qfloat8_e4m3fn",
600
+ "activations": "none"
601
+ },
602
+ "transformer_blocks.10.attn.to_q": {
603
+ "weights": "qfloat8_e4m3fn",
604
+ "activations": "none"
605
+ },
606
+ "transformer_blocks.10.attn.to_k": {
607
+ "weights": "qfloat8_e4m3fn",
608
+ "activations": "none"
609
+ },
610
+ "transformer_blocks.10.attn.to_v": {
611
+ "weights": "qfloat8_e4m3fn",
612
+ "activations": "none"
613
+ },
614
+ "transformer_blocks.10.attn.add_k_proj": {
615
+ "weights": "qfloat8_e4m3fn",
616
+ "activations": "none"
617
+ },
618
+ "transformer_blocks.10.attn.add_v_proj": {
619
+ "weights": "qfloat8_e4m3fn",
620
+ "activations": "none"
621
+ },
622
+ "transformer_blocks.10.attn.add_q_proj": {
623
+ "weights": "qfloat8_e4m3fn",
624
+ "activations": "none"
625
+ },
626
+ "transformer_blocks.10.attn.to_out.0": {
627
+ "weights": "qfloat8_e4m3fn",
628
+ "activations": "none"
629
+ },
630
+ "transformer_blocks.10.attn.to_add_out": {
631
+ "weights": "qfloat8_e4m3fn",
632
+ "activations": "none"
633
+ },
634
+ "transformer_blocks.10.ff.net.0.proj": {
635
+ "weights": "qfloat8_e4m3fn",
636
+ "activations": "none"
637
+ },
638
+ "transformer_blocks.10.ff.net.2": {
639
+ "weights": "qfloat8_e4m3fn",
640
+ "activations": "none"
641
+ },
642
+ "transformer_blocks.10.ff_context.net.0.proj": {
643
+ "weights": "qfloat8_e4m3fn",
644
+ "activations": "none"
645
+ },
646
+ "transformer_blocks.10.ff_context.net.2": {
647
+ "weights": "qfloat8_e4m3fn",
648
+ "activations": "none"
649
+ },
650
+ "transformer_blocks.11.norm1.linear": {
651
+ "weights": "qfloat8_e4m3fn",
652
+ "activations": "none"
653
+ },
654
+ "transformer_blocks.11.norm1_context.linear": {
655
+ "weights": "qfloat8_e4m3fn",
656
+ "activations": "none"
657
+ },
658
+ "transformer_blocks.11.attn.to_q": {
659
+ "weights": "qfloat8_e4m3fn",
660
+ "activations": "none"
661
+ },
662
+ "transformer_blocks.11.attn.to_k": {
663
+ "weights": "qfloat8_e4m3fn",
664
+ "activations": "none"
665
+ },
666
+ "transformer_blocks.11.attn.to_v": {
667
+ "weights": "qfloat8_e4m3fn",
668
+ "activations": "none"
669
+ },
670
+ "transformer_blocks.11.attn.add_k_proj": {
671
+ "weights": "qfloat8_e4m3fn",
672
+ "activations": "none"
673
+ },
674
+ "transformer_blocks.11.attn.add_v_proj": {
675
+ "weights": "qfloat8_e4m3fn",
676
+ "activations": "none"
677
+ },
678
+ "transformer_blocks.11.attn.add_q_proj": {
679
+ "weights": "qfloat8_e4m3fn",
680
+ "activations": "none"
681
+ },
682
+ "transformer_blocks.11.attn.to_out.0": {
683
+ "weights": "qfloat8_e4m3fn",
684
+ "activations": "none"
685
+ },
686
+ "transformer_blocks.11.attn.to_add_out": {
687
+ "weights": "qfloat8_e4m3fn",
688
+ "activations": "none"
689
+ },
690
+ "transformer_blocks.11.ff.net.0.proj": {
691
+ "weights": "qfloat8_e4m3fn",
692
+ "activations": "none"
693
+ },
694
+ "transformer_blocks.11.ff.net.2": {
695
+ "weights": "qfloat8_e4m3fn",
696
+ "activations": "none"
697
+ },
698
+ "transformer_blocks.11.ff_context.net.0.proj": {
699
+ "weights": "qfloat8_e4m3fn",
700
+ "activations": "none"
701
+ },
702
+ "transformer_blocks.11.ff_context.net.2": {
703
+ "weights": "qfloat8_e4m3fn",
704
+ "activations": "none"
705
+ },
706
+ "transformer_blocks.12.norm1.linear": {
707
+ "weights": "qfloat8_e4m3fn",
708
+ "activations": "none"
709
+ },
710
+ "transformer_blocks.12.norm1_context.linear": {
711
+ "weights": "qfloat8_e4m3fn",
712
+ "activations": "none"
713
+ },
714
+ "transformer_blocks.12.attn.to_q": {
715
+ "weights": "qfloat8_e4m3fn",
716
+ "activations": "none"
717
+ },
718
+ "transformer_blocks.12.attn.to_k": {
719
+ "weights": "qfloat8_e4m3fn",
720
+ "activations": "none"
721
+ },
722
+ "transformer_blocks.12.attn.to_v": {
723
+ "weights": "qfloat8_e4m3fn",
724
+ "activations": "none"
725
+ },
726
+ "transformer_blocks.12.attn.add_k_proj": {
727
+ "weights": "qfloat8_e4m3fn",
728
+ "activations": "none"
729
+ },
730
+ "transformer_blocks.12.attn.add_v_proj": {
731
+ "weights": "qfloat8_e4m3fn",
732
+ "activations": "none"
733
+ },
734
+ "transformer_blocks.12.attn.add_q_proj": {
735
+ "weights": "qfloat8_e4m3fn",
736
+ "activations": "none"
737
+ },
738
+ "transformer_blocks.12.attn.to_out.0": {
739
+ "weights": "qfloat8_e4m3fn",
740
+ "activations": "none"
741
+ },
742
+ "transformer_blocks.12.attn.to_add_out": {
743
+ "weights": "qfloat8_e4m3fn",
744
+ "activations": "none"
745
+ },
746
+ "transformer_blocks.12.ff.net.0.proj": {
747
+ "weights": "qfloat8_e4m3fn",
748
+ "activations": "none"
749
+ },
750
+ "transformer_blocks.12.ff.net.2": {
751
+ "weights": "qfloat8_e4m3fn",
752
+ "activations": "none"
753
+ },
754
+ "transformer_blocks.12.ff_context.net.0.proj": {
755
+ "weights": "qfloat8_e4m3fn",
756
+ "activations": "none"
757
+ },
758
+ "transformer_blocks.12.ff_context.net.2": {
759
+ "weights": "qfloat8_e4m3fn",
760
+ "activations": "none"
761
+ },
762
+ "transformer_blocks.13.norm1.linear": {
763
+ "weights": "qfloat8_e4m3fn",
764
+ "activations": "none"
765
+ },
766
+ "transformer_blocks.13.norm1_context.linear": {
767
+ "weights": "qfloat8_e4m3fn",
768
+ "activations": "none"
769
+ },
770
+ "transformer_blocks.13.attn.to_q": {
771
+ "weights": "qfloat8_e4m3fn",
772
+ "activations": "none"
773
+ },
774
+ "transformer_blocks.13.attn.to_k": {
775
+ "weights": "qfloat8_e4m3fn",
776
+ "activations": "none"
777
+ },
778
+ "transformer_blocks.13.attn.to_v": {
779
+ "weights": "qfloat8_e4m3fn",
780
+ "activations": "none"
781
+ },
782
+ "transformer_blocks.13.attn.add_k_proj": {
783
+ "weights": "qfloat8_e4m3fn",
784
+ "activations": "none"
785
+ },
786
+ "transformer_blocks.13.attn.add_v_proj": {
787
+ "weights": "qfloat8_e4m3fn",
788
+ "activations": "none"
789
+ },
790
+ "transformer_blocks.13.attn.add_q_proj": {
791
+ "weights": "qfloat8_e4m3fn",
792
+ "activations": "none"
793
+ },
794
+ "transformer_blocks.13.attn.to_out.0": {
795
+ "weights": "qfloat8_e4m3fn",
796
+ "activations": "none"
797
+ },
798
+ "transformer_blocks.13.attn.to_add_out": {
799
+ "weights": "qfloat8_e4m3fn",
800
+ "activations": "none"
801
+ },
802
+ "transformer_blocks.13.ff.net.0.proj": {
803
+ "weights": "qfloat8_e4m3fn",
804
+ "activations": "none"
805
+ },
806
+ "transformer_blocks.13.ff.net.2": {
807
+ "weights": "qfloat8_e4m3fn",
808
+ "activations": "none"
809
+ },
810
+ "transformer_blocks.13.ff_context.net.0.proj": {
811
+ "weights": "qfloat8_e4m3fn",
812
+ "activations": "none"
813
+ },
814
+ "transformer_blocks.13.ff_context.net.2": {
815
+ "weights": "qfloat8_e4m3fn",
816
+ "activations": "none"
817
+ },
818
+ "transformer_blocks.14.norm1.linear": {
819
+ "weights": "qfloat8_e4m3fn",
820
+ "activations": "none"
821
+ },
822
+ "transformer_blocks.14.norm1_context.linear": {
823
+ "weights": "qfloat8_e4m3fn",
824
+ "activations": "none"
825
+ },
826
+ "transformer_blocks.14.attn.to_q": {
827
+ "weights": "qfloat8_e4m3fn",
828
+ "activations": "none"
829
+ },
830
+ "transformer_blocks.14.attn.to_k": {
831
+ "weights": "qfloat8_e4m3fn",
832
+ "activations": "none"
833
+ },
834
+ "transformer_blocks.14.attn.to_v": {
835
+ "weights": "qfloat8_e4m3fn",
836
+ "activations": "none"
837
+ },
838
+ "transformer_blocks.14.attn.add_k_proj": {
839
+ "weights": "qfloat8_e4m3fn",
840
+ "activations": "none"
841
+ },
842
+ "transformer_blocks.14.attn.add_v_proj": {
843
+ "weights": "qfloat8_e4m3fn",
844
+ "activations": "none"
845
+ },
846
+ "transformer_blocks.14.attn.add_q_proj": {
847
+ "weights": "qfloat8_e4m3fn",
848
+ "activations": "none"
849
+ },
850
+ "transformer_blocks.14.attn.to_out.0": {
851
+ "weights": "qfloat8_e4m3fn",
852
+ "activations": "none"
853
+ },
854
+ "transformer_blocks.14.attn.to_add_out": {
855
+ "weights": "qfloat8_e4m3fn",
856
+ "activations": "none"
857
+ },
858
+ "transformer_blocks.14.ff.net.0.proj": {
859
+ "weights": "qfloat8_e4m3fn",
860
+ "activations": "none"
861
+ },
862
+ "transformer_blocks.14.ff.net.2": {
863
+ "weights": "qfloat8_e4m3fn",
864
+ "activations": "none"
865
+ },
866
+ "transformer_blocks.14.ff_context.net.0.proj": {
867
+ "weights": "qfloat8_e4m3fn",
868
+ "activations": "none"
869
+ },
870
+ "transformer_blocks.14.ff_context.net.2": {
871
+ "weights": "qfloat8_e4m3fn",
872
+ "activations": "none"
873
+ },
874
+ "transformer_blocks.15.norm1.linear": {
875
+ "weights": "qfloat8_e4m3fn",
876
+ "activations": "none"
877
+ },
878
+ "transformer_blocks.15.norm1_context.linear": {
879
+ "weights": "qfloat8_e4m3fn",
880
+ "activations": "none"
881
+ },
882
+ "transformer_blocks.15.attn.to_q": {
883
+ "weights": "qfloat8_e4m3fn",
884
+ "activations": "none"
885
+ },
886
+ "transformer_blocks.15.attn.to_k": {
887
+ "weights": "qfloat8_e4m3fn",
888
+ "activations": "none"
889
+ },
890
+ "transformer_blocks.15.attn.to_v": {
891
+ "weights": "qfloat8_e4m3fn",
892
+ "activations": "none"
893
+ },
894
+ "transformer_blocks.15.attn.add_k_proj": {
895
+ "weights": "qfloat8_e4m3fn",
896
+ "activations": "none"
897
+ },
898
+ "transformer_blocks.15.attn.add_v_proj": {
899
+ "weights": "qfloat8_e4m3fn",
900
+ "activations": "none"
901
+ },
902
+ "transformer_blocks.15.attn.add_q_proj": {
903
+ "weights": "qfloat8_e4m3fn",
904
+ "activations": "none"
905
+ },
906
+ "transformer_blocks.15.attn.to_out.0": {
907
+ "weights": "qfloat8_e4m3fn",
908
+ "activations": "none"
909
+ },
910
+ "transformer_blocks.15.attn.to_add_out": {
911
+ "weights": "qfloat8_e4m3fn",
912
+ "activations": "none"
913
+ },
914
+ "transformer_blocks.15.ff.net.0.proj": {
915
+ "weights": "qfloat8_e4m3fn",
916
+ "activations": "none"
917
+ },
918
+ "transformer_blocks.15.ff.net.2": {
919
+ "weights": "qfloat8_e4m3fn",
920
+ "activations": "none"
921
+ },
922
+ "transformer_blocks.15.ff_context.net.0.proj": {
923
+ "weights": "qfloat8_e4m3fn",
924
+ "activations": "none"
925
+ },
926
+ "transformer_blocks.15.ff_context.net.2": {
927
+ "weights": "qfloat8_e4m3fn",
928
+ "activations": "none"
929
+ },
930
+ "transformer_blocks.16.norm1.linear": {
931
+ "weights": "qfloat8_e4m3fn",
932
+ "activations": "none"
933
+ },
934
+ "transformer_blocks.16.norm1_context.linear": {
935
+ "weights": "qfloat8_e4m3fn",
936
+ "activations": "none"
937
+ },
938
+ "transformer_blocks.16.attn.to_q": {
939
+ "weights": "qfloat8_e4m3fn",
940
+ "activations": "none"
941
+ },
942
+ "transformer_blocks.16.attn.to_k": {
943
+ "weights": "qfloat8_e4m3fn",
944
+ "activations": "none"
945
+ },
946
+ "transformer_blocks.16.attn.to_v": {
947
+ "weights": "qfloat8_e4m3fn",
948
+ "activations": "none"
949
+ },
950
+ "transformer_blocks.16.attn.add_k_proj": {
951
+ "weights": "qfloat8_e4m3fn",
952
+ "activations": "none"
953
+ },
954
+ "transformer_blocks.16.attn.add_v_proj": {
955
+ "weights": "qfloat8_e4m3fn",
956
+ "activations": "none"
957
+ },
958
+ "transformer_blocks.16.attn.add_q_proj": {
959
+ "weights": "qfloat8_e4m3fn",
960
+ "activations": "none"
961
+ },
962
+ "transformer_blocks.16.attn.to_out.0": {
963
+ "weights": "qfloat8_e4m3fn",
964
+ "activations": "none"
965
+ },
966
+ "transformer_blocks.16.attn.to_add_out": {
967
+ "weights": "qfloat8_e4m3fn",
968
+ "activations": "none"
969
+ },
970
+ "transformer_blocks.16.ff.net.0.proj": {
971
+ "weights": "qfloat8_e4m3fn",
972
+ "activations": "none"
973
+ },
974
+ "transformer_blocks.16.ff.net.2": {
975
+ "weights": "qfloat8_e4m3fn",
976
+ "activations": "none"
977
+ },
978
+ "transformer_blocks.16.ff_context.net.0.proj": {
979
+ "weights": "qfloat8_e4m3fn",
980
+ "activations": "none"
981
+ },
982
+ "transformer_blocks.16.ff_context.net.2": {
983
+ "weights": "qfloat8_e4m3fn",
984
+ "activations": "none"
985
+ },
986
+ "transformer_blocks.17.norm1.linear": {
987
+ "weights": "qfloat8_e4m3fn",
988
+ "activations": "none"
989
+ },
990
+ "transformer_blocks.17.norm1_context.linear": {
991
+ "weights": "qfloat8_e4m3fn",
992
+ "activations": "none"
993
+ },
994
+ "transformer_blocks.17.attn.to_q": {
995
+ "weights": "qfloat8_e4m3fn",
996
+ "activations": "none"
997
+ },
998
+ "transformer_blocks.17.attn.to_k": {
999
+ "weights": "qfloat8_e4m3fn",
1000
+ "activations": "none"
1001
+ },
1002
+ "transformer_blocks.17.attn.to_v": {
1003
+ "weights": "qfloat8_e4m3fn",
1004
+ "activations": "none"
1005
+ },
1006
+ "transformer_blocks.17.attn.add_k_proj": {
1007
+ "weights": "qfloat8_e4m3fn",
1008
+ "activations": "none"
1009
+ },
1010
+ "transformer_blocks.17.attn.add_v_proj": {
1011
+ "weights": "qfloat8_e4m3fn",
1012
+ "activations": "none"
1013
+ },
1014
+ "transformer_blocks.17.attn.add_q_proj": {
1015
+ "weights": "qfloat8_e4m3fn",
1016
+ "activations": "none"
1017
+ },
1018
+ "transformer_blocks.17.attn.to_out.0": {
1019
+ "weights": "qfloat8_e4m3fn",
1020
+ "activations": "none"
1021
+ },
1022
+ "transformer_blocks.17.attn.to_add_out": {
1023
+ "weights": "qfloat8_e4m3fn",
1024
+ "activations": "none"
1025
+ },
1026
+ "transformer_blocks.17.ff.net.0.proj": {
1027
+ "weights": "qfloat8_e4m3fn",
1028
+ "activations": "none"
1029
+ },
1030
+ "transformer_blocks.17.ff.net.2": {
1031
+ "weights": "qfloat8_e4m3fn",
1032
+ "activations": "none"
1033
+ },
1034
+ "transformer_blocks.17.ff_context.net.0.proj": {
1035
+ "weights": "qfloat8_e4m3fn",
1036
+ "activations": "none"
1037
+ },
1038
+ "transformer_blocks.17.ff_context.net.2": {
1039
+ "weights": "qfloat8_e4m3fn",
1040
+ "activations": "none"
1041
+ },
1042
+ "transformer_blocks.18.norm1.linear": {
1043
+ "weights": "qfloat8_e4m3fn",
1044
+ "activations": "none"
1045
+ },
1046
+ "transformer_blocks.18.norm1_context.linear": {
1047
+ "weights": "qfloat8_e4m3fn",
1048
+ "activations": "none"
1049
+ },
1050
+ "transformer_blocks.18.attn.to_q": {
1051
+ "weights": "qfloat8_e4m3fn",
1052
+ "activations": "none"
1053
+ },
1054
+ "transformer_blocks.18.attn.to_k": {
1055
+ "weights": "qfloat8_e4m3fn",
1056
+ "activations": "none"
1057
+ },
1058
+ "transformer_blocks.18.attn.to_v": {
1059
+ "weights": "qfloat8_e4m3fn",
1060
+ "activations": "none"
1061
+ },
1062
+ "transformer_blocks.18.attn.add_k_proj": {
1063
+ "weights": "qfloat8_e4m3fn",
1064
+ "activations": "none"
1065
+ },
1066
+ "transformer_blocks.18.attn.add_v_proj": {
1067
+ "weights": "qfloat8_e4m3fn",
1068
+ "activations": "none"
1069
+ },
1070
+ "transformer_blocks.18.attn.add_q_proj": {
1071
+ "weights": "qfloat8_e4m3fn",
1072
+ "activations": "none"
1073
+ },
1074
+ "transformer_blocks.18.attn.to_out.0": {
1075
+ "weights": "qfloat8_e4m3fn",
1076
+ "activations": "none"
1077
+ },
1078
+ "transformer_blocks.18.attn.to_add_out": {
1079
+ "weights": "qfloat8_e4m3fn",
1080
+ "activations": "none"
1081
+ },
1082
+ "transformer_blocks.18.ff.net.0.proj": {
1083
+ "weights": "qfloat8_e4m3fn",
1084
+ "activations": "none"
1085
+ },
1086
+ "transformer_blocks.18.ff.net.2": {
1087
+ "weights": "qfloat8_e4m3fn",
1088
+ "activations": "none"
1089
+ },
1090
+ "transformer_blocks.18.ff_context.net.0.proj": {
1091
+ "weights": "qfloat8_e4m3fn",
1092
+ "activations": "none"
1093
+ },
1094
+ "transformer_blocks.18.ff_context.net.2": {
1095
+ "weights": "qfloat8_e4m3fn",
1096
+ "activations": "none"
1097
+ },
1098
+ "single_transformer_blocks.0.norm.linear": {
1099
+ "weights": "qfloat8_e4m3fn",
1100
+ "activations": "none"
1101
+ },
1102
+ "single_transformer_blocks.0.proj_mlp": {
1103
+ "weights": "qfloat8_e4m3fn",
1104
+ "activations": "none"
1105
+ },
1106
+ "single_transformer_blocks.0.proj_out": {
1107
+ "weights": "qfloat8_e4m3fn",
1108
+ "activations": "none"
1109
+ },
1110
+ "single_transformer_blocks.0.attn.to_q": {
1111
+ "weights": "qfloat8_e4m3fn",
1112
+ "activations": "none"
1113
+ },
1114
+ "single_transformer_blocks.0.attn.to_k": {
1115
+ "weights": "qfloat8_e4m3fn",
1116
+ "activations": "none"
1117
+ },
1118
+ "single_transformer_blocks.0.attn.to_v": {
1119
+ "weights": "qfloat8_e4m3fn",
1120
+ "activations": "none"
1121
+ },
1122
+ "single_transformer_blocks.1.norm.linear": {
1123
+ "weights": "qfloat8_e4m3fn",
1124
+ "activations": "none"
1125
+ },
1126
+ "single_transformer_blocks.1.proj_mlp": {
1127
+ "weights": "qfloat8_e4m3fn",
1128
+ "activations": "none"
1129
+ },
1130
+ "single_transformer_blocks.1.proj_out": {
1131
+ "weights": "qfloat8_e4m3fn",
1132
+ "activations": "none"
1133
+ },
1134
+ "single_transformer_blocks.1.attn.to_q": {
1135
+ "weights": "qfloat8_e4m3fn",
1136
+ "activations": "none"
1137
+ },
1138
+ "single_transformer_blocks.1.attn.to_k": {
1139
+ "weights": "qfloat8_e4m3fn",
1140
+ "activations": "none"
1141
+ },
1142
+ "single_transformer_blocks.1.attn.to_v": {
1143
+ "weights": "qfloat8_e4m3fn",
1144
+ "activations": "none"
1145
+ },
1146
+ "single_transformer_blocks.2.norm.linear": {
1147
+ "weights": "qfloat8_e4m3fn",
1148
+ "activations": "none"
1149
+ },
1150
+ "single_transformer_blocks.2.proj_mlp": {
1151
+ "weights": "qfloat8_e4m3fn",
1152
+ "activations": "none"
1153
+ },
1154
+ "single_transformer_blocks.2.proj_out": {
1155
+ "weights": "qfloat8_e4m3fn",
1156
+ "activations": "none"
1157
+ },
1158
+ "single_transformer_blocks.2.attn.to_q": {
1159
+ "weights": "qfloat8_e4m3fn",
1160
+ "activations": "none"
1161
+ },
1162
+ "single_transformer_blocks.2.attn.to_k": {
1163
+ "weights": "qfloat8_e4m3fn",
1164
+ "activations": "none"
1165
+ },
1166
+ "single_transformer_blocks.2.attn.to_v": {
1167
+ "weights": "qfloat8_e4m3fn",
1168
+ "activations": "none"
1169
+ },
1170
+ "single_transformer_blocks.3.norm.linear": {
1171
+ "weights": "qfloat8_e4m3fn",
1172
+ "activations": "none"
1173
+ },
1174
+ "single_transformer_blocks.3.proj_mlp": {
1175
+ "weights": "qfloat8_e4m3fn",
1176
+ "activations": "none"
1177
+ },
1178
+ "single_transformer_blocks.3.proj_out": {
1179
+ "weights": "qfloat8_e4m3fn",
1180
+ "activations": "none"
1181
+ },
1182
+ "single_transformer_blocks.3.attn.to_q": {
1183
+ "weights": "qfloat8_e4m3fn",
1184
+ "activations": "none"
1185
+ },
1186
+ "single_transformer_blocks.3.attn.to_k": {
1187
+ "weights": "qfloat8_e4m3fn",
1188
+ "activations": "none"
1189
+ },
1190
+ "single_transformer_blocks.3.attn.to_v": {
1191
+ "weights": "qfloat8_e4m3fn",
1192
+ "activations": "none"
1193
+ },
1194
+ "single_transformer_blocks.4.norm.linear": {
1195
+ "weights": "qfloat8_e4m3fn",
1196
+ "activations": "none"
1197
+ },
1198
+ "single_transformer_blocks.4.proj_mlp": {
1199
+ "weights": "qfloat8_e4m3fn",
1200
+ "activations": "none"
1201
+ },
1202
+ "single_transformer_blocks.4.proj_out": {
1203
+ "weights": "qfloat8_e4m3fn",
1204
+ "activations": "none"
1205
+ },
1206
+ "single_transformer_blocks.4.attn.to_q": {
1207
+ "weights": "qfloat8_e4m3fn",
1208
+ "activations": "none"
1209
+ },
1210
+ "single_transformer_blocks.4.attn.to_k": {
1211
+ "weights": "qfloat8_e4m3fn",
1212
+ "activations": "none"
1213
+ },
1214
+ "single_transformer_blocks.4.attn.to_v": {
1215
+ "weights": "qfloat8_e4m3fn",
1216
+ "activations": "none"
1217
+ },
1218
+ "single_transformer_blocks.5.norm.linear": {
1219
+ "weights": "qfloat8_e4m3fn",
1220
+ "activations": "none"
1221
+ },
1222
+ "single_transformer_blocks.5.proj_mlp": {
1223
+ "weights": "qfloat8_e4m3fn",
1224
+ "activations": "none"
1225
+ },
1226
+ "single_transformer_blocks.5.proj_out": {
1227
+ "weights": "qfloat8_e4m3fn",
1228
+ "activations": "none"
1229
+ },
1230
+ "single_transformer_blocks.5.attn.to_q": {
1231
+ "weights": "qfloat8_e4m3fn",
1232
+ "activations": "none"
1233
+ },
1234
+ "single_transformer_blocks.5.attn.to_k": {
1235
+ "weights": "qfloat8_e4m3fn",
1236
+ "activations": "none"
1237
+ },
1238
+ "single_transformer_blocks.5.attn.to_v": {
1239
+ "weights": "qfloat8_e4m3fn",
1240
+ "activations": "none"
1241
+ },
1242
+ "single_transformer_blocks.6.norm.linear": {
1243
+ "weights": "qfloat8_e4m3fn",
1244
+ "activations": "none"
1245
+ },
1246
+ "single_transformer_blocks.6.proj_mlp": {
1247
+ "weights": "qfloat8_e4m3fn",
1248
+ "activations": "none"
1249
+ },
1250
+ "single_transformer_blocks.6.proj_out": {
1251
+ "weights": "qfloat8_e4m3fn",
1252
+ "activations": "none"
1253
+ },
1254
+ "single_transformer_blocks.6.attn.to_q": {
1255
+ "weights": "qfloat8_e4m3fn",
1256
+ "activations": "none"
1257
+ },
1258
+ "single_transformer_blocks.6.attn.to_k": {
1259
+ "weights": "qfloat8_e4m3fn",
1260
+ "activations": "none"
1261
+ },
1262
+ "single_transformer_blocks.6.attn.to_v": {
1263
+ "weights": "qfloat8_e4m3fn",
1264
+ "activations": "none"
1265
+ },
1266
+ "single_transformer_blocks.7.norm.linear": {
1267
+ "weights": "qfloat8_e4m3fn",
1268
+ "activations": "none"
1269
+ },
1270
+ "single_transformer_blocks.7.proj_mlp": {
1271
+ "weights": "qfloat8_e4m3fn",
1272
+ "activations": "none"
1273
+ },
1274
+ "single_transformer_blocks.7.proj_out": {
1275
+ "weights": "qfloat8_e4m3fn",
1276
+ "activations": "none"
1277
+ },
1278
+ "single_transformer_blocks.7.attn.to_q": {
1279
+ "weights": "qfloat8_e4m3fn",
1280
+ "activations": "none"
1281
+ },
1282
+ "single_transformer_blocks.7.attn.to_k": {
1283
+ "weights": "qfloat8_e4m3fn",
1284
+ "activations": "none"
1285
+ },
1286
+ "single_transformer_blocks.7.attn.to_v": {
1287
+ "weights": "qfloat8_e4m3fn",
1288
+ "activations": "none"
1289
+ },
1290
+ "single_transformer_blocks.8.norm.linear": {
1291
+ "weights": "qfloat8_e4m3fn",
1292
+ "activations": "none"
1293
+ },
1294
+ "single_transformer_blocks.8.proj_mlp": {
1295
+ "weights": "qfloat8_e4m3fn",
1296
+ "activations": "none"
1297
+ },
1298
+ "single_transformer_blocks.8.proj_out": {
1299
+ "weights": "qfloat8_e4m3fn",
1300
+ "activations": "none"
1301
+ },
1302
+ "single_transformer_blocks.8.attn.to_q": {
1303
+ "weights": "qfloat8_e4m3fn",
1304
+ "activations": "none"
1305
+ },
1306
+ "single_transformer_blocks.8.attn.to_k": {
1307
+ "weights": "qfloat8_e4m3fn",
1308
+ "activations": "none"
1309
+ },
1310
+ "single_transformer_blocks.8.attn.to_v": {
1311
+ "weights": "qfloat8_e4m3fn",
1312
+ "activations": "none"
1313
+ },
1314
+ "single_transformer_blocks.9.norm.linear": {
1315
+ "weights": "qfloat8_e4m3fn",
1316
+ "activations": "none"
1317
+ },
1318
+ "single_transformer_blocks.9.proj_mlp": {
1319
+ "weights": "qfloat8_e4m3fn",
1320
+ "activations": "none"
1321
+ },
1322
+ "single_transformer_blocks.9.proj_out": {
1323
+ "weights": "qfloat8_e4m3fn",
1324
+ "activations": "none"
1325
+ },
1326
+ "single_transformer_blocks.9.attn.to_q": {
1327
+ "weights": "qfloat8_e4m3fn",
1328
+ "activations": "none"
1329
+ },
1330
+ "single_transformer_blocks.9.attn.to_k": {
1331
+ "weights": "qfloat8_e4m3fn",
1332
+ "activations": "none"
1333
+ },
1334
+ "single_transformer_blocks.9.attn.to_v": {
1335
+ "weights": "qfloat8_e4m3fn",
1336
+ "activations": "none"
1337
+ },
1338
+ "single_transformer_blocks.10.norm.linear": {
1339
+ "weights": "qfloat8_e4m3fn",
1340
+ "activations": "none"
1341
+ },
1342
+ "single_transformer_blocks.10.proj_mlp": {
1343
+ "weights": "qfloat8_e4m3fn",
1344
+ "activations": "none"
1345
+ },
1346
+ "single_transformer_blocks.10.proj_out": {
1347
+ "weights": "qfloat8_e4m3fn",
1348
+ "activations": "none"
1349
+ },
1350
+ "single_transformer_blocks.10.attn.to_q": {
1351
+ "weights": "qfloat8_e4m3fn",
1352
+ "activations": "none"
1353
+ },
1354
+ "single_transformer_blocks.10.attn.to_k": {
1355
+ "weights": "qfloat8_e4m3fn",
1356
+ "activations": "none"
1357
+ },
1358
+ "single_transformer_blocks.10.attn.to_v": {
1359
+ "weights": "qfloat8_e4m3fn",
1360
+ "activations": "none"
1361
+ },
1362
+ "single_transformer_blocks.11.norm.linear": {
1363
+ "weights": "qfloat8_e4m3fn",
1364
+ "activations": "none"
1365
+ },
1366
+ "single_transformer_blocks.11.proj_mlp": {
1367
+ "weights": "qfloat8_e4m3fn",
1368
+ "activations": "none"
1369
+ },
1370
+ "single_transformer_blocks.11.proj_out": {
1371
+ "weights": "qfloat8_e4m3fn",
1372
+ "activations": "none"
1373
+ },
1374
+ "single_transformer_blocks.11.attn.to_q": {
1375
+ "weights": "qfloat8_e4m3fn",
1376
+ "activations": "none"
1377
+ },
1378
+ "single_transformer_blocks.11.attn.to_k": {
1379
+ "weights": "qfloat8_e4m3fn",
1380
+ "activations": "none"
1381
+ },
1382
+ "single_transformer_blocks.11.attn.to_v": {
1383
+ "weights": "qfloat8_e4m3fn",
1384
+ "activations": "none"
1385
+ },
1386
+ "single_transformer_blocks.12.norm.linear": {
1387
+ "weights": "qfloat8_e4m3fn",
1388
+ "activations": "none"
1389
+ },
1390
+ "single_transformer_blocks.12.proj_mlp": {
1391
+ "weights": "qfloat8_e4m3fn",
1392
+ "activations": "none"
1393
+ },
1394
+ "single_transformer_blocks.12.proj_out": {
1395
+ "weights": "qfloat8_e4m3fn",
1396
+ "activations": "none"
1397
+ },
1398
+ "single_transformer_blocks.12.attn.to_q": {
1399
+ "weights": "qfloat8_e4m3fn",
1400
+ "activations": "none"
1401
+ },
1402
+ "single_transformer_blocks.12.attn.to_k": {
1403
+ "weights": "qfloat8_e4m3fn",
1404
+ "activations": "none"
1405
+ },
1406
+ "single_transformer_blocks.12.attn.to_v": {
1407
+ "weights": "qfloat8_e4m3fn",
1408
+ "activations": "none"
1409
+ },
1410
+ "single_transformer_blocks.13.norm.linear": {
1411
+ "weights": "qfloat8_e4m3fn",
1412
+ "activations": "none"
1413
+ },
1414
+ "single_transformer_blocks.13.proj_mlp": {
1415
+ "weights": "qfloat8_e4m3fn",
1416
+ "activations": "none"
1417
+ },
1418
+ "single_transformer_blocks.13.proj_out": {
1419
+ "weights": "qfloat8_e4m3fn",
1420
+ "activations": "none"
1421
+ },
1422
+ "single_transformer_blocks.13.attn.to_q": {
1423
+ "weights": "qfloat8_e4m3fn",
1424
+ "activations": "none"
1425
+ },
1426
+ "single_transformer_blocks.13.attn.to_k": {
1427
+ "weights": "qfloat8_e4m3fn",
1428
+ "activations": "none"
1429
+ },
1430
+ "single_transformer_blocks.13.attn.to_v": {
1431
+ "weights": "qfloat8_e4m3fn",
1432
+ "activations": "none"
1433
+ },
1434
+ "single_transformer_blocks.14.norm.linear": {
1435
+ "weights": "qfloat8_e4m3fn",
1436
+ "activations": "none"
1437
+ },
1438
+ "single_transformer_blocks.14.proj_mlp": {
1439
+ "weights": "qfloat8_e4m3fn",
1440
+ "activations": "none"
1441
+ },
1442
+ "single_transformer_blocks.14.proj_out": {
1443
+ "weights": "qfloat8_e4m3fn",
1444
+ "activations": "none"
1445
+ },
1446
+ "single_transformer_blocks.14.attn.to_q": {
1447
+ "weights": "qfloat8_e4m3fn",
1448
+ "activations": "none"
1449
+ },
1450
+ "single_transformer_blocks.14.attn.to_k": {
1451
+ "weights": "qfloat8_e4m3fn",
1452
+ "activations": "none"
1453
+ },
1454
+ "single_transformer_blocks.14.attn.to_v": {
1455
+ "weights": "qfloat8_e4m3fn",
1456
+ "activations": "none"
1457
+ },
1458
+ "single_transformer_blocks.15.norm.linear": {
1459
+ "weights": "qfloat8_e4m3fn",
1460
+ "activations": "none"
1461
+ },
1462
+ "single_transformer_blocks.15.proj_mlp": {
1463
+ "weights": "qfloat8_e4m3fn",
1464
+ "activations": "none"
1465
+ },
1466
+ "single_transformer_blocks.15.proj_out": {
1467
+ "weights": "qfloat8_e4m3fn",
1468
+ "activations": "none"
1469
+ },
1470
+ "single_transformer_blocks.15.attn.to_q": {
1471
+ "weights": "qfloat8_e4m3fn",
1472
+ "activations": "none"
1473
+ },
1474
+ "single_transformer_blocks.15.attn.to_k": {
1475
+ "weights": "qfloat8_e4m3fn",
1476
+ "activations": "none"
1477
+ },
1478
+ "single_transformer_blocks.15.attn.to_v": {
1479
+ "weights": "qfloat8_e4m3fn",
1480
+ "activations": "none"
1481
+ },
1482
+ "single_transformer_blocks.16.norm.linear": {
1483
+ "weights": "qfloat8_e4m3fn",
1484
+ "activations": "none"
1485
+ },
1486
+ "single_transformer_blocks.16.proj_mlp": {
1487
+ "weights": "qfloat8_e4m3fn",
1488
+ "activations": "none"
1489
+ },
1490
+ "single_transformer_blocks.16.proj_out": {
1491
+ "weights": "qfloat8_e4m3fn",
1492
+ "activations": "none"
1493
+ },
1494
+ "single_transformer_blocks.16.attn.to_q": {
1495
+ "weights": "qfloat8_e4m3fn",
1496
+ "activations": "none"
1497
+ },
1498
+ "single_transformer_blocks.16.attn.to_k": {
1499
+ "weights": "qfloat8_e4m3fn",
1500
+ "activations": "none"
1501
+ },
1502
+ "single_transformer_blocks.16.attn.to_v": {
1503
+ "weights": "qfloat8_e4m3fn",
1504
+ "activations": "none"
1505
+ },
1506
+ "single_transformer_blocks.17.norm.linear": {
1507
+ "weights": "qfloat8_e4m3fn",
1508
+ "activations": "none"
1509
+ },
1510
+ "single_transformer_blocks.17.proj_mlp": {
1511
+ "weights": "qfloat8_e4m3fn",
1512
+ "activations": "none"
1513
+ },
1514
+ "single_transformer_blocks.17.proj_out": {
1515
+ "weights": "qfloat8_e4m3fn",
1516
+ "activations": "none"
1517
+ },
1518
+ "single_transformer_blocks.17.attn.to_q": {
1519
+ "weights": "qfloat8_e4m3fn",
1520
+ "activations": "none"
1521
+ },
1522
+ "single_transformer_blocks.17.attn.to_k": {
1523
+ "weights": "qfloat8_e4m3fn",
1524
+ "activations": "none"
1525
+ },
1526
+ "single_transformer_blocks.17.attn.to_v": {
1527
+ "weights": "qfloat8_e4m3fn",
1528
+ "activations": "none"
1529
+ },
1530
+ "single_transformer_blocks.18.norm.linear": {
1531
+ "weights": "qfloat8_e4m3fn",
1532
+ "activations": "none"
1533
+ },
1534
+ "single_transformer_blocks.18.proj_mlp": {
1535
+ "weights": "qfloat8_e4m3fn",
1536
+ "activations": "none"
1537
+ },
1538
+ "single_transformer_blocks.18.proj_out": {
1539
+ "weights": "qfloat8_e4m3fn",
1540
+ "activations": "none"
1541
+ },
1542
+ "single_transformer_blocks.18.attn.to_q": {
1543
+ "weights": "qfloat8_e4m3fn",
1544
+ "activations": "none"
1545
+ },
1546
+ "single_transformer_blocks.18.attn.to_k": {
1547
+ "weights": "qfloat8_e4m3fn",
1548
+ "activations": "none"
1549
+ },
1550
+ "single_transformer_blocks.18.attn.to_v": {
1551
+ "weights": "qfloat8_e4m3fn",
1552
+ "activations": "none"
1553
+ },
1554
+ "single_transformer_blocks.19.norm.linear": {
1555
+ "weights": "qfloat8_e4m3fn",
1556
+ "activations": "none"
1557
+ },
1558
+ "single_transformer_blocks.19.proj_mlp": {
1559
+ "weights": "qfloat8_e4m3fn",
1560
+ "activations": "none"
1561
+ },
1562
+ "single_transformer_blocks.19.proj_out": {
1563
+ "weights": "qfloat8_e4m3fn",
1564
+ "activations": "none"
1565
+ },
1566
+ "single_transformer_blocks.19.attn.to_q": {
1567
+ "weights": "qfloat8_e4m3fn",
1568
+ "activations": "none"
1569
+ },
1570
+ "single_transformer_blocks.19.attn.to_k": {
1571
+ "weights": "qfloat8_e4m3fn",
1572
+ "activations": "none"
1573
+ },
1574
+ "single_transformer_blocks.19.attn.to_v": {
1575
+ "weights": "qfloat8_e4m3fn",
1576
+ "activations": "none"
1577
+ },
1578
+ "single_transformer_blocks.20.norm.linear": {
1579
+ "weights": "qfloat8_e4m3fn",
1580
+ "activations": "none"
1581
+ },
1582
+ "single_transformer_blocks.20.proj_mlp": {
1583
+ "weights": "qfloat8_e4m3fn",
1584
+ "activations": "none"
1585
+ },
1586
+ "single_transformer_blocks.20.proj_out": {
1587
+ "weights": "qfloat8_e4m3fn",
1588
+ "activations": "none"
1589
+ },
1590
+ "single_transformer_blocks.20.attn.to_q": {
1591
+ "weights": "qfloat8_e4m3fn",
1592
+ "activations": "none"
1593
+ },
1594
+ "single_transformer_blocks.20.attn.to_k": {
1595
+ "weights": "qfloat8_e4m3fn",
1596
+ "activations": "none"
1597
+ },
1598
+ "single_transformer_blocks.20.attn.to_v": {
1599
+ "weights": "qfloat8_e4m3fn",
1600
+ "activations": "none"
1601
+ },
1602
+ "single_transformer_blocks.21.norm.linear": {
1603
+ "weights": "qfloat8_e4m3fn",
1604
+ "activations": "none"
1605
+ },
1606
+ "single_transformer_blocks.21.proj_mlp": {
1607
+ "weights": "qfloat8_e4m3fn",
1608
+ "activations": "none"
1609
+ },
1610
+ "single_transformer_blocks.21.proj_out": {
1611
+ "weights": "qfloat8_e4m3fn",
1612
+ "activations": "none"
1613
+ },
1614
+ "single_transformer_blocks.21.attn.to_q": {
1615
+ "weights": "qfloat8_e4m3fn",
1616
+ "activations": "none"
1617
+ },
1618
+ "single_transformer_blocks.21.attn.to_k": {
1619
+ "weights": "qfloat8_e4m3fn",
1620
+ "activations": "none"
1621
+ },
1622
+ "single_transformer_blocks.21.attn.to_v": {
1623
+ "weights": "qfloat8_e4m3fn",
1624
+ "activations": "none"
1625
+ },
1626
+ "single_transformer_blocks.22.norm.linear": {
1627
+ "weights": "qfloat8_e4m3fn",
1628
+ "activations": "none"
1629
+ },
1630
+ "single_transformer_blocks.22.proj_mlp": {
1631
+ "weights": "qfloat8_e4m3fn",
1632
+ "activations": "none"
1633
+ },
1634
+ "single_transformer_blocks.22.proj_out": {
1635
+ "weights": "qfloat8_e4m3fn",
1636
+ "activations": "none"
1637
+ },
1638
+ "single_transformer_blocks.22.attn.to_q": {
1639
+ "weights": "qfloat8_e4m3fn",
1640
+ "activations": "none"
1641
+ },
1642
+ "single_transformer_blocks.22.attn.to_k": {
1643
+ "weights": "qfloat8_e4m3fn",
1644
+ "activations": "none"
1645
+ },
1646
+ "single_transformer_blocks.22.attn.to_v": {
1647
+ "weights": "qfloat8_e4m3fn",
1648
+ "activations": "none"
1649
+ },
1650
+ "single_transformer_blocks.23.norm.linear": {
1651
+ "weights": "qfloat8_e4m3fn",
1652
+ "activations": "none"
1653
+ },
1654
+ "single_transformer_blocks.23.proj_mlp": {
1655
+ "weights": "qfloat8_e4m3fn",
1656
+ "activations": "none"
1657
+ },
1658
+ "single_transformer_blocks.23.proj_out": {
1659
+ "weights": "qfloat8_e4m3fn",
1660
+ "activations": "none"
1661
+ },
1662
+ "single_transformer_blocks.23.attn.to_q": {
1663
+ "weights": "qfloat8_e4m3fn",
1664
+ "activations": "none"
1665
+ },
1666
+ "single_transformer_blocks.23.attn.to_k": {
1667
+ "weights": "qfloat8_e4m3fn",
1668
+ "activations": "none"
1669
+ },
1670
+ "single_transformer_blocks.23.attn.to_v": {
1671
+ "weights": "qfloat8_e4m3fn",
1672
+ "activations": "none"
1673
+ },
1674
+ "single_transformer_blocks.24.norm.linear": {
1675
+ "weights": "qfloat8_e4m3fn",
1676
+ "activations": "none"
1677
+ },
1678
+ "single_transformer_blocks.24.proj_mlp": {
1679
+ "weights": "qfloat8_e4m3fn",
1680
+ "activations": "none"
1681
+ },
1682
+ "single_transformer_blocks.24.proj_out": {
1683
+ "weights": "qfloat8_e4m3fn",
1684
+ "activations": "none"
1685
+ },
1686
+ "single_transformer_blocks.24.attn.to_q": {
1687
+ "weights": "qfloat8_e4m3fn",
1688
+ "activations": "none"
1689
+ },
1690
+ "single_transformer_blocks.24.attn.to_k": {
1691
+ "weights": "qfloat8_e4m3fn",
1692
+ "activations": "none"
1693
+ },
1694
+ "single_transformer_blocks.24.attn.to_v": {
1695
+ "weights": "qfloat8_e4m3fn",
1696
+ "activations": "none"
1697
+ },
1698
+ "single_transformer_blocks.25.norm.linear": {
1699
+ "weights": "qfloat8_e4m3fn",
1700
+ "activations": "none"
1701
+ },
1702
+ "single_transformer_blocks.25.proj_mlp": {
1703
+ "weights": "qfloat8_e4m3fn",
1704
+ "activations": "none"
1705
+ },
1706
+ "single_transformer_blocks.25.proj_out": {
1707
+ "weights": "qfloat8_e4m3fn",
1708
+ "activations": "none"
1709
+ },
1710
+ "single_transformer_blocks.25.attn.to_q": {
1711
+ "weights": "qfloat8_e4m3fn",
1712
+ "activations": "none"
1713
+ },
1714
+ "single_transformer_blocks.25.attn.to_k": {
1715
+ "weights": "qfloat8_e4m3fn",
1716
+ "activations": "none"
1717
+ },
1718
+ "single_transformer_blocks.25.attn.to_v": {
1719
+ "weights": "qfloat8_e4m3fn",
1720
+ "activations": "none"
1721
+ },
1722
+ "single_transformer_blocks.26.norm.linear": {
1723
+ "weights": "qfloat8_e4m3fn",
1724
+ "activations": "none"
1725
+ },
1726
+ "single_transformer_blocks.26.proj_mlp": {
1727
+ "weights": "qfloat8_e4m3fn",
1728
+ "activations": "none"
1729
+ },
1730
+ "single_transformer_blocks.26.proj_out": {
1731
+ "weights": "qfloat8_e4m3fn",
1732
+ "activations": "none"
1733
+ },
1734
+ "single_transformer_blocks.26.attn.to_q": {
1735
+ "weights": "qfloat8_e4m3fn",
1736
+ "activations": "none"
1737
+ },
1738
+ "single_transformer_blocks.26.attn.to_k": {
1739
+ "weights": "qfloat8_e4m3fn",
1740
+ "activations": "none"
1741
+ },
1742
+ "single_transformer_blocks.26.attn.to_v": {
1743
+ "weights": "qfloat8_e4m3fn",
1744
+ "activations": "none"
1745
+ },
1746
+ "single_transformer_blocks.27.norm.linear": {
1747
+ "weights": "qfloat8_e4m3fn",
1748
+ "activations": "none"
1749
+ },
1750
+ "single_transformer_blocks.27.proj_mlp": {
1751
+ "weights": "qfloat8_e4m3fn",
1752
+ "activations": "none"
1753
+ },
1754
+ "single_transformer_blocks.27.proj_out": {
1755
+ "weights": "qfloat8_e4m3fn",
1756
+ "activations": "none"
1757
+ },
1758
+ "single_transformer_blocks.27.attn.to_q": {
1759
+ "weights": "qfloat8_e4m3fn",
1760
+ "activations": "none"
1761
+ },
1762
+ "single_transformer_blocks.27.attn.to_k": {
1763
+ "weights": "qfloat8_e4m3fn",
1764
+ "activations": "none"
1765
+ },
1766
+ "single_transformer_blocks.27.attn.to_v": {
1767
+ "weights": "qfloat8_e4m3fn",
1768
+ "activations": "none"
1769
+ },
1770
+ "single_transformer_blocks.28.norm.linear": {
1771
+ "weights": "qfloat8_e4m3fn",
1772
+ "activations": "none"
1773
+ },
1774
+ "single_transformer_blocks.28.proj_mlp": {
1775
+ "weights": "qfloat8_e4m3fn",
1776
+ "activations": "none"
1777
+ },
1778
+ "single_transformer_blocks.28.proj_out": {
1779
+ "weights": "qfloat8_e4m3fn",
1780
+ "activations": "none"
1781
+ },
1782
+ "single_transformer_blocks.28.attn.to_q": {
1783
+ "weights": "qfloat8_e4m3fn",
1784
+ "activations": "none"
1785
+ },
1786
+ "single_transformer_blocks.28.attn.to_k": {
1787
+ "weights": "qfloat8_e4m3fn",
1788
+ "activations": "none"
1789
+ },
1790
+ "single_transformer_blocks.28.attn.to_v": {
1791
+ "weights": "qfloat8_e4m3fn",
1792
+ "activations": "none"
1793
+ },
1794
+ "single_transformer_blocks.29.norm.linear": {
1795
+ "weights": "qfloat8_e4m3fn",
1796
+ "activations": "none"
1797
+ },
1798
+ "single_transformer_blocks.29.proj_mlp": {
1799
+ "weights": "qfloat8_e4m3fn",
1800
+ "activations": "none"
1801
+ },
1802
+ "single_transformer_blocks.29.proj_out": {
1803
+ "weights": "qfloat8_e4m3fn",
1804
+ "activations": "none"
1805
+ },
1806
+ "single_transformer_blocks.29.attn.to_q": {
1807
+ "weights": "qfloat8_e4m3fn",
1808
+ "activations": "none"
1809
+ },
1810
+ "single_transformer_blocks.29.attn.to_k": {
1811
+ "weights": "qfloat8_e4m3fn",
1812
+ "activations": "none"
1813
+ },
1814
+ "single_transformer_blocks.29.attn.to_v": {
1815
+ "weights": "qfloat8_e4m3fn",
1816
+ "activations": "none"
1817
+ },
1818
+ "single_transformer_blocks.30.norm.linear": {
1819
+ "weights": "qfloat8_e4m3fn",
1820
+ "activations": "none"
1821
+ },
1822
+ "single_transformer_blocks.30.proj_mlp": {
1823
+ "weights": "qfloat8_e4m3fn",
1824
+ "activations": "none"
1825
+ },
1826
+ "single_transformer_blocks.30.proj_out": {
1827
+ "weights": "qfloat8_e4m3fn",
1828
+ "activations": "none"
1829
+ },
1830
+ "single_transformer_blocks.30.attn.to_q": {
1831
+ "weights": "qfloat8_e4m3fn",
1832
+ "activations": "none"
1833
+ },
1834
+ "single_transformer_blocks.30.attn.to_k": {
1835
+ "weights": "qfloat8_e4m3fn",
1836
+ "activations": "none"
1837
+ },
1838
+ "single_transformer_blocks.30.attn.to_v": {
1839
+ "weights": "qfloat8_e4m3fn",
1840
+ "activations": "none"
1841
+ },
1842
+ "single_transformer_blocks.31.norm.linear": {
1843
+ "weights": "qfloat8_e4m3fn",
1844
+ "activations": "none"
1845
+ },
1846
+ "single_transformer_blocks.31.proj_mlp": {
1847
+ "weights": "qfloat8_e4m3fn",
1848
+ "activations": "none"
1849
+ },
1850
+ "single_transformer_blocks.31.proj_out": {
1851
+ "weights": "qfloat8_e4m3fn",
1852
+ "activations": "none"
1853
+ },
1854
+ "single_transformer_blocks.31.attn.to_q": {
1855
+ "weights": "qfloat8_e4m3fn",
1856
+ "activations": "none"
1857
+ },
1858
+ "single_transformer_blocks.31.attn.to_k": {
1859
+ "weights": "qfloat8_e4m3fn",
1860
+ "activations": "none"
1861
+ },
1862
+ "single_transformer_blocks.31.attn.to_v": {
1863
+ "weights": "qfloat8_e4m3fn",
1864
+ "activations": "none"
1865
+ },
1866
+ "single_transformer_blocks.32.norm.linear": {
1867
+ "weights": "qfloat8_e4m3fn",
1868
+ "activations": "none"
1869
+ },
1870
+ "single_transformer_blocks.32.proj_mlp": {
1871
+ "weights": "qfloat8_e4m3fn",
1872
+ "activations": "none"
1873
+ },
1874
+ "single_transformer_blocks.32.proj_out": {
1875
+ "weights": "qfloat8_e4m3fn",
1876
+ "activations": "none"
1877
+ },
1878
+ "single_transformer_blocks.32.attn.to_q": {
1879
+ "weights": "qfloat8_e4m3fn",
1880
+ "activations": "none"
1881
+ },
1882
+ "single_transformer_blocks.32.attn.to_k": {
1883
+ "weights": "qfloat8_e4m3fn",
1884
+ "activations": "none"
1885
+ },
1886
+ "single_transformer_blocks.32.attn.to_v": {
1887
+ "weights": "qfloat8_e4m3fn",
1888
+ "activations": "none"
1889
+ },
1890
+ "single_transformer_blocks.33.norm.linear": {
1891
+ "weights": "qfloat8_e4m3fn",
1892
+ "activations": "none"
1893
+ },
1894
+ "single_transformer_blocks.33.proj_mlp": {
1895
+ "weights": "qfloat8_e4m3fn",
1896
+ "activations": "none"
1897
+ },
1898
+ "single_transformer_blocks.33.proj_out": {
1899
+ "weights": "qfloat8_e4m3fn",
1900
+ "activations": "none"
1901
+ },
1902
+ "single_transformer_blocks.33.attn.to_q": {
1903
+ "weights": "qfloat8_e4m3fn",
1904
+ "activations": "none"
1905
+ },
1906
+ "single_transformer_blocks.33.attn.to_k": {
1907
+ "weights": "qfloat8_e4m3fn",
1908
+ "activations": "none"
1909
+ },
1910
+ "single_transformer_blocks.33.attn.to_v": {
1911
+ "weights": "qfloat8_e4m3fn",
1912
+ "activations": "none"
1913
+ },
1914
+ "single_transformer_blocks.34.norm.linear": {
1915
+ "weights": "qfloat8_e4m3fn",
1916
+ "activations": "none"
1917
+ },
1918
+ "single_transformer_blocks.34.proj_mlp": {
1919
+ "weights": "qfloat8_e4m3fn",
1920
+ "activations": "none"
1921
+ },
1922
+ "single_transformer_blocks.34.proj_out": {
1923
+ "weights": "qfloat8_e4m3fn",
1924
+ "activations": "none"
1925
+ },
1926
+ "single_transformer_blocks.34.attn.to_q": {
1927
+ "weights": "qfloat8_e4m3fn",
1928
+ "activations": "none"
1929
+ },
1930
+ "single_transformer_blocks.34.attn.to_k": {
1931
+ "weights": "qfloat8_e4m3fn",
1932
+ "activations": "none"
1933
+ },
1934
+ "single_transformer_blocks.34.attn.to_v": {
1935
+ "weights": "qfloat8_e4m3fn",
1936
+ "activations": "none"
1937
+ },
1938
+ "single_transformer_blocks.35.norm.linear": {
1939
+ "weights": "qfloat8_e4m3fn",
1940
+ "activations": "none"
1941
+ },
1942
+ "single_transformer_blocks.35.proj_mlp": {
1943
+ "weights": "qfloat8_e4m3fn",
1944
+ "activations": "none"
1945
+ },
1946
+ "single_transformer_blocks.35.proj_out": {
1947
+ "weights": "qfloat8_e4m3fn",
1948
+ "activations": "none"
1949
+ },
1950
+ "single_transformer_blocks.35.attn.to_q": {
1951
+ "weights": "qfloat8_e4m3fn",
1952
+ "activations": "none"
1953
+ },
1954
+ "single_transformer_blocks.35.attn.to_k": {
1955
+ "weights": "qfloat8_e4m3fn",
1956
+ "activations": "none"
1957
+ },
1958
+ "single_transformer_blocks.35.attn.to_v": {
1959
+ "weights": "qfloat8_e4m3fn",
1960
+ "activations": "none"
1961
+ },
1962
+ "single_transformer_blocks.36.norm.linear": {
1963
+ "weights": "qfloat8_e4m3fn",
1964
+ "activations": "none"
1965
+ },
1966
+ "single_transformer_blocks.36.proj_mlp": {
1967
+ "weights": "qfloat8_e4m3fn",
1968
+ "activations": "none"
1969
+ },
1970
+ "single_transformer_blocks.36.proj_out": {
1971
+ "weights": "qfloat8_e4m3fn",
1972
+ "activations": "none"
1973
+ },
1974
+ "single_transformer_blocks.36.attn.to_q": {
1975
+ "weights": "qfloat8_e4m3fn",
1976
+ "activations": "none"
1977
+ },
1978
+ "single_transformer_blocks.36.attn.to_k": {
1979
+ "weights": "qfloat8_e4m3fn",
1980
+ "activations": "none"
1981
+ },
1982
+ "single_transformer_blocks.36.attn.to_v": {
1983
+ "weights": "qfloat8_e4m3fn",
1984
+ "activations": "none"
1985
+ },
1986
+ "single_transformer_blocks.37.norm.linear": {
1987
+ "weights": "qfloat8_e4m3fn",
1988
+ "activations": "none"
1989
+ },
1990
+ "single_transformer_blocks.37.proj_mlp": {
1991
+ "weights": "qfloat8_e4m3fn",
1992
+ "activations": "none"
1993
+ },
1994
+ "single_transformer_blocks.37.proj_out": {
1995
+ "weights": "qfloat8_e4m3fn",
1996
+ "activations": "none"
1997
+ },
1998
+ "single_transformer_blocks.37.attn.to_q": {
1999
+ "weights": "qfloat8_e4m3fn",
2000
+ "activations": "none"
2001
+ },
2002
+ "single_transformer_blocks.37.attn.to_k": {
2003
+ "weights": "qfloat8_e4m3fn",
2004
+ "activations": "none"
2005
+ },
2006
+ "single_transformer_blocks.37.attn.to_v": {
2007
+ "weights": "qfloat8_e4m3fn",
2008
+ "activations": "none"
2009
+ },
2010
+ "norm_out.linear": {
2011
+ "weights": "qfloat8_e4m3fn",
2012
+ "activations": "none"
2013
+ },
2014
+ "proj_out": {
2015
+ "weights": "qfloat8_e4m3fn",
2016
+ "activations": "none"
2017
+ }
2018
+ }