dn6 HF staff commited on
Commit
97ea675
1 Parent(s): 89a09a4

Update Diffusers T5 weights

Browse files
.huggingface/.gitignore ADDED
@@ -0,0 +1 @@
 
 
1
+ *
.huggingface/download/.gitattributes.lock ADDED
File without changes
.huggingface/download/.gitattributes.metadata ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ 1d5ece72b189f7322a5405e9682711ffe9580378
2
+ aafb9f7bec760b90674d44433d49b8ea46d6cb9e
3
+ 1730831164.6254377
.huggingface/download/README.md.lock ADDED
File without changes
.huggingface/download/README.md.metadata ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ 1d5ece72b189f7322a5405e9682711ffe9580378
2
+ 30ae5c954ea1067c62d7fc3447832e17512eb2c2
3
+ 1730831164.8530679
.huggingface/download/assets/grid.gif.lock ADDED
File without changes
.huggingface/download/assets/grid.gif.metadata ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ 1d5ece72b189f7322a5405e9682711ffe9580378
2
+ 85588586e345d5302112486b34a59c34240ab759c3048f0abc3868d8f3fdaa5d
3
+ 1730831165.5925496
.huggingface/download/decoder.safetensors.lock ADDED
File without changes
.huggingface/download/decoder.safetensors.metadata ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ 1d5ece72b189f7322a5405e9682711ffe9580378
2
+ 641920faaf20e5404ddb5553ce3e295c21ed9b4bc5f6fe7c930811b84099cb14
3
+ 1730831176.4544442
.huggingface/download/dit.safetensors.lock ADDED
File without changes
.huggingface/download/dit.safetensors.metadata ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ 1d5ece72b189f7322a5405e9682711ffe9580378
2
+ 5c273a0d4fbb0de835e31b7cb5228089ba0ee54d4c3a087e81d354cef3d8bbf1
3
+ 1730831387.0177512
.huggingface/download/encoder.safetensors.lock ADDED
File without changes
.huggingface/download/encoder.safetensors.metadata ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ 1d5ece72b189f7322a5405e9682711ffe9580378
2
+ d3a8827a66b58a479d97420a9bf77e59078d88f538298469d8db28c37bd556ae
3
+ 1730831389.0487227
README.md CHANGED
@@ -27,7 +27,7 @@ Install using [uv](https://github.com/astral-sh/uv):
27
 
28
  ```bash
29
  git clone https://github.com/genmoai/models
30
- cd models
31
  pip install uv
32
  uv venv .venv
33
  source .venv/bin/activate
@@ -133,7 +133,7 @@ pipe.enable_vae_tiling()
133
  prompt = "Close-up of a chameleon's eye, with its scaly skin changing color. Ultra high resolution 4k."
134
 
135
  with torch.autocast("cuda", torch.bfloat16, cache_enabled=False):
136
- frames = pipe(prompt, num_frames=84).frames[0]
137
 
138
  export_to_video(frames, "mochi.mp4", fps=30)
139
  ```
@@ -154,7 +154,7 @@ pipe.enable_model_cpu_offload()
154
  pipe.enable_vae_tiling()
155
 
156
  prompt = "Close-up of a chameleon's eye, with its scaly skin changing color. Ultra high resolution 4k."
157
- frames = pipe(prompt, num_frames=84).frames[0]
158
 
159
  export_to_video(frames, "mochi.mp4", fps=30)
160
  ```
@@ -164,20 +164,20 @@ To learn more check out the [Diffusers](https://huggingface.co/docs/diffusers/ma
164
 
165
  ## Model Architecture
166
 
167
- Mochi 1 represents a significant advancement in open-source video generation, featuring a 10 billion parameter diffusion model built on our novel Asymmetric Diffusion Transformer (AsymmDiT) architecture. Trained entirely from scratch, it is the largest video generative model ever openly released. And best of all, it’s a simple, hackable architecture. Additionally, we are releasing an inference harness that includes an efficient context parallel implementation.
168
 
169
- Alongside Mochi, we are open-sourcing our video AsymmVAE. We use an asymmetric encoder-decoder structure to build an efficient high quality compression model. Our AsymmVAE causally compresses videos to a 128x smaller size, with an 8x8 spatial and a 6x temporal compression to a 12-channel latent space.
170
 
171
  ### AsymmVAE Model Specs
172
- |Params <br> Count | Enc Base <br> Channels | Dec Base <br> Channels |Latent <br> Dim | Spatial <br> Compression | Temporal <br> Compression |
173
  |:--:|:--:|:--:|:--:|:--:|:--:|
174
- |362M | 64 | 128 | 12 | 8x8 | 6x |
175
 
176
  An AsymmDiT efficiently processes user prompts alongside compressed video tokens by streamlining text processing and focusing neural network capacity on visual reasoning. AsymmDiT jointly attends to text and visual tokens with multi-modal self-attention and learns separate MLP layers for each modality, similar to Stable Diffusion 3. However, our visual stream has nearly 4 times as many parameters as the text stream via a larger hidden dimension. To unify the modalities in self-attention, we use non-square QKV and output projection layers. This asymmetric design reduces inference memory requirements.
177
  Many modern diffusion models use multiple pretrained language models to represent user prompts. In contrast, Mochi 1 simply encodes prompts with a single T5-XXL language model.
178
 
179
  ### AsymmDiT Model Specs
180
- |Params <br> Count | Num <br> Layers | Num <br> Heads | Visual <br> Dim | Text <br> Dim | Visual <br> Tokens | Text <br> Tokens |
181
  |:--:|:--:|:--:|:--:|:--:|:--:|:--:|
182
  |10B | 48 | 24 | 3072 | 1536 | 44520 | 256 |
183
 
 
27
 
28
  ```bash
29
  git clone https://github.com/genmoai/models
30
+ cd models
31
  pip install uv
32
  uv venv .venv
33
  source .venv/bin/activate
 
133
  prompt = "Close-up of a chameleon's eye, with its scaly skin changing color. Ultra high resolution 4k."
134
 
135
  with torch.autocast("cuda", torch.bfloat16, cache_enabled=False):
136
+ frames = pipe(prompt, num_frames=85).frames[0]
137
 
138
  export_to_video(frames, "mochi.mp4", fps=30)
139
  ```
 
154
  pipe.enable_vae_tiling()
155
 
156
  prompt = "Close-up of a chameleon's eye, with its scaly skin changing color. Ultra high resolution 4k."
157
+ frames = pipe(prompt, num_frames=85).frames[0]
158
 
159
  export_to_video(frames, "mochi.mp4", fps=30)
160
  ```
 
164
 
165
  ## Model Architecture
166
 
167
+ Mochi 1 represents a significant advancement in open-source video generation, featuring a 10 billion parameter diffusion model built on our novel Asymmetric Diffusion Transformer (AsymmDiT) architecture. Trained entirely from scratch, it is the largest video generative model ever openly released. And best of all, it’s a simple, hackable architecture. Additionally, we are releasing an inference harness that includes an efficient context parallel implementation.
168
 
169
+ Alongside Mochi, we are open-sourcing our video AsymmVAE. We use an asymmetric encoder-decoder structure to build an efficient high quality compression model. Our AsymmVAE causally compresses videos to a 128x smaller size, with an 8x8 spatial and a 6x temporal compression to a 12-channel latent space.
170
 
171
  ### AsymmVAE Model Specs
172
+ |Params <br> Count | Enc Base <br> Channels | Dec Base <br> Channels |Latent <br> Dim | Spatial <br> Compression | Temporal <br> Compression |
173
  |:--:|:--:|:--:|:--:|:--:|:--:|
174
+ |362M | 64 | 128 | 12 | 8x8 | 6x |
175
 
176
  An AsymmDiT efficiently processes user prompts alongside compressed video tokens by streamlining text processing and focusing neural network capacity on visual reasoning. AsymmDiT jointly attends to text and visual tokens with multi-modal self-attention and learns separate MLP layers for each modality, similar to Stable Diffusion 3. However, our visual stream has nearly 4 times as many parameters as the text stream via a larger hidden dimension. To unify the modalities in self-attention, we use non-square QKV and output projection layers. This asymmetric design reduces inference memory requirements.
177
  Many modern diffusion models use multiple pretrained language models to represent user prompts. In contrast, Mochi 1 simply encodes prompts with a single T5-XXL language model.
178
 
179
  ### AsymmDiT Model Specs
180
+ |Params <br> Count | Num <br> Layers | Num <br> Heads | Visual <br> Dim | Text <br> Dim | Visual <br> Tokens | Text <br> Tokens |
181
  |:--:|:--:|:--:|:--:|:--:|:--:|:--:|
182
  |10B | 48 | 24 | 3072 | 1536 | 44520 | 256 |
183
 
text_encoder/config.json CHANGED
@@ -1,4 +1,5 @@
1
  {
 
2
  "architectures": [
3
  "T5EncoderModel"
4
  ],
@@ -24,7 +25,7 @@
24
  "relative_attention_max_distance": 128,
25
  "relative_attention_num_buckets": 32,
26
  "tie_word_embeddings": false,
27
- "torch_dtype": "bfloat16",
28
  "transformers_version": "4.45.2",
29
  "use_cache": true,
30
  "vocab_size": 32128
 
1
  {
2
+ "_name_or_path": "google/t5-v1_1-xxl",
3
  "architectures": [
4
  "T5EncoderModel"
5
  ],
 
25
  "relative_attention_max_distance": 128,
26
  "relative_attention_num_buckets": 32,
27
  "tie_word_embeddings": false,
28
+ "torch_dtype": "float32",
29
  "transformers_version": "4.45.2",
30
  "use_cache": true,
31
  "vocab_size": 32128
text_encoder/model-00001-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a68b2c8c080696a10109612a649bc69330991ecfea65930ccfdfbdb011f2686
3
+ size 4989319680
text_encoder/model-00002-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b8ed6556d7507e38af5b428c605fb2a6f2bdb7e80bd481308b865f7a40c551ca
3
+ size 4999830656
text_encoder/model-00003-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c831635f83041f83faf0024b39c6ecb21b45d70dd38a63ea5bac6c7c6e5e558c
3
+ size 4865612720
text_encoder/model-00004-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:02a5f2d69205be92ad48fe5d712d38c2ff55627969116aeffc58bd75a28da468
3
+ size 4194506688
text_encoder/model.safetensors.index.json CHANGED
@@ -1,226 +1,226 @@
1
  {
2
  "metadata": {
3
- "total_size": 9524621312
4
  },
5
  "weight_map": {
6
- "encoder.block.0.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
7
- "encoder.block.0.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
8
- "encoder.block.0.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
9
- "encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight": "model-00001-of-00002.safetensors",
10
- "encoder.block.0.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
11
- "encoder.block.0.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
12
- "encoder.block.0.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
13
- "encoder.block.0.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
14
- "encoder.block.0.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
15
- "encoder.block.0.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
16
- "encoder.block.1.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
17
- "encoder.block.1.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
18
- "encoder.block.1.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
19
- "encoder.block.1.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
20
- "encoder.block.1.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
21
- "encoder.block.1.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
22
- "encoder.block.1.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
23
- "encoder.block.1.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
24
- "encoder.block.1.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
25
- "encoder.block.10.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
26
- "encoder.block.10.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
27
- "encoder.block.10.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
28
- "encoder.block.10.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
29
- "encoder.block.10.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
30
- "encoder.block.10.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
31
- "encoder.block.10.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
32
- "encoder.block.10.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
33
- "encoder.block.10.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
34
- "encoder.block.11.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
35
- "encoder.block.11.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
36
- "encoder.block.11.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
37
- "encoder.block.11.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
38
- "encoder.block.11.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
39
- "encoder.block.11.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
40
- "encoder.block.11.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
41
- "encoder.block.11.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
42
- "encoder.block.11.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
43
- "encoder.block.12.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
44
- "encoder.block.12.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
45
- "encoder.block.12.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
46
- "encoder.block.12.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
47
- "encoder.block.12.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
48
- "encoder.block.12.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
49
- "encoder.block.12.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
50
- "encoder.block.12.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
51
- "encoder.block.12.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
52
- "encoder.block.13.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
53
- "encoder.block.13.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
54
- "encoder.block.13.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
55
- "encoder.block.13.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
56
- "encoder.block.13.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
57
- "encoder.block.13.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
58
- "encoder.block.13.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
59
- "encoder.block.13.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
60
- "encoder.block.13.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
61
- "encoder.block.14.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
62
- "encoder.block.14.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
63
- "encoder.block.14.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
64
- "encoder.block.14.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
65
- "encoder.block.14.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
66
- "encoder.block.14.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
67
- "encoder.block.14.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
68
- "encoder.block.14.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
69
- "encoder.block.14.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
70
- "encoder.block.15.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
71
- "encoder.block.15.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
72
- "encoder.block.15.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
73
- "encoder.block.15.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
74
- "encoder.block.15.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
75
- "encoder.block.15.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
76
- "encoder.block.15.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
77
- "encoder.block.15.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
78
- "encoder.block.15.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
79
- "encoder.block.16.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
80
- "encoder.block.16.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
81
- "encoder.block.16.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
82
- "encoder.block.16.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
83
- "encoder.block.16.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
84
- "encoder.block.16.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
85
- "encoder.block.16.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
86
- "encoder.block.16.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
87
- "encoder.block.16.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
88
- "encoder.block.17.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
89
- "encoder.block.17.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
90
- "encoder.block.17.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
91
- "encoder.block.17.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
92
- "encoder.block.17.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
93
- "encoder.block.17.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
94
- "encoder.block.17.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
95
- "encoder.block.17.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
96
- "encoder.block.17.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
97
- "encoder.block.18.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
98
- "encoder.block.18.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
99
- "encoder.block.18.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
100
- "encoder.block.18.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
101
- "encoder.block.18.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
102
- "encoder.block.18.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
103
- "encoder.block.18.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
104
- "encoder.block.18.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
105
- "encoder.block.18.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
106
- "encoder.block.19.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
107
- "encoder.block.19.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
108
- "encoder.block.19.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
109
- "encoder.block.19.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
110
- "encoder.block.19.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
111
- "encoder.block.19.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
112
- "encoder.block.19.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
113
- "encoder.block.19.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
114
- "encoder.block.19.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
115
- "encoder.block.2.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
116
- "encoder.block.2.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
117
- "encoder.block.2.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
118
- "encoder.block.2.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
119
- "encoder.block.2.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
120
- "encoder.block.2.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
121
- "encoder.block.2.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
122
- "encoder.block.2.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
123
- "encoder.block.2.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
124
- "encoder.block.20.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
125
- "encoder.block.20.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
126
- "encoder.block.20.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
127
- "encoder.block.20.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
128
- "encoder.block.20.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
129
- "encoder.block.20.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
130
- "encoder.block.20.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
131
- "encoder.block.20.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
132
- "encoder.block.20.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
133
- "encoder.block.21.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
134
- "encoder.block.21.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
135
- "encoder.block.21.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
136
- "encoder.block.21.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
137
- "encoder.block.21.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
138
- "encoder.block.21.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
139
- "encoder.block.21.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
140
- "encoder.block.21.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
141
- "encoder.block.21.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
142
- "encoder.block.22.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
143
- "encoder.block.22.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
144
- "encoder.block.22.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
145
- "encoder.block.22.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
146
- "encoder.block.22.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
147
- "encoder.block.22.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
148
- "encoder.block.22.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
149
- "encoder.block.22.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
150
- "encoder.block.22.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
151
- "encoder.block.23.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
152
- "encoder.block.23.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
153
- "encoder.block.23.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
154
- "encoder.block.23.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
155
- "encoder.block.23.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
156
- "encoder.block.23.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
157
- "encoder.block.23.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
158
- "encoder.block.23.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
159
- "encoder.block.23.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
160
- "encoder.block.3.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
161
- "encoder.block.3.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
162
- "encoder.block.3.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
163
- "encoder.block.3.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
164
- "encoder.block.3.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
165
- "encoder.block.3.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
166
- "encoder.block.3.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
167
- "encoder.block.3.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
168
- "encoder.block.3.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
169
- "encoder.block.4.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
170
- "encoder.block.4.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
171
- "encoder.block.4.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
172
- "encoder.block.4.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
173
- "encoder.block.4.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
174
- "encoder.block.4.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
175
- "encoder.block.4.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
176
- "encoder.block.4.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
177
- "encoder.block.4.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
178
- "encoder.block.5.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
179
- "encoder.block.5.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
180
- "encoder.block.5.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
181
- "encoder.block.5.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
182
- "encoder.block.5.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
183
- "encoder.block.5.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
184
- "encoder.block.5.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
185
- "encoder.block.5.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
186
- "encoder.block.5.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
187
- "encoder.block.6.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
188
- "encoder.block.6.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
189
- "encoder.block.6.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
190
- "encoder.block.6.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
191
- "encoder.block.6.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
192
- "encoder.block.6.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
193
- "encoder.block.6.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
194
- "encoder.block.6.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
195
- "encoder.block.6.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
196
- "encoder.block.7.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
197
- "encoder.block.7.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
198
- "encoder.block.7.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
199
- "encoder.block.7.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
200
- "encoder.block.7.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
201
- "encoder.block.7.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
202
- "encoder.block.7.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
203
- "encoder.block.7.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
204
- "encoder.block.7.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
205
- "encoder.block.8.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
206
- "encoder.block.8.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
207
- "encoder.block.8.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
208
- "encoder.block.8.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
209
- "encoder.block.8.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
210
- "encoder.block.8.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
211
- "encoder.block.8.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
212
- "encoder.block.8.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
213
- "encoder.block.8.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
214
- "encoder.block.9.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
215
- "encoder.block.9.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
216
- "encoder.block.9.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
217
- "encoder.block.9.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
218
- "encoder.block.9.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
219
- "encoder.block.9.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
220
- "encoder.block.9.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
221
- "encoder.block.9.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
222
- "encoder.block.9.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
223
- "encoder.final_layer_norm.weight": "model-00002-of-00002.safetensors",
224
- "shared.weight": "model-00001-of-00002.safetensors"
225
  }
226
  }
 
1
  {
2
  "metadata": {
3
+ "total_size": 19049242624
4
  },
5
  "weight_map": {
6
+ "encoder.block.0.layer.0.SelfAttention.k.weight": "model-00001-of-00004.safetensors",
7
+ "encoder.block.0.layer.0.SelfAttention.o.weight": "model-00001-of-00004.safetensors",
8
+ "encoder.block.0.layer.0.SelfAttention.q.weight": "model-00001-of-00004.safetensors",
9
+ "encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight": "model-00001-of-00004.safetensors",
10
+ "encoder.block.0.layer.0.SelfAttention.v.weight": "model-00001-of-00004.safetensors",
11
+ "encoder.block.0.layer.0.layer_norm.weight": "model-00001-of-00004.safetensors",
12
+ "encoder.block.0.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00004.safetensors",
13
+ "encoder.block.0.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00004.safetensors",
14
+ "encoder.block.0.layer.1.DenseReluDense.wo.weight": "model-00001-of-00004.safetensors",
15
+ "encoder.block.0.layer.1.layer_norm.weight": "model-00001-of-00004.safetensors",
16
+ "encoder.block.1.layer.0.SelfAttention.k.weight": "model-00001-of-00004.safetensors",
17
+ "encoder.block.1.layer.0.SelfAttention.o.weight": "model-00001-of-00004.safetensors",
18
+ "encoder.block.1.layer.0.SelfAttention.q.weight": "model-00001-of-00004.safetensors",
19
+ "encoder.block.1.layer.0.SelfAttention.v.weight": "model-00001-of-00004.safetensors",
20
+ "encoder.block.1.layer.0.layer_norm.weight": "model-00001-of-00004.safetensors",
21
+ "encoder.block.1.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00004.safetensors",
22
+ "encoder.block.1.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00004.safetensors",
23
+ "encoder.block.1.layer.1.DenseReluDense.wo.weight": "model-00001-of-00004.safetensors",
24
+ "encoder.block.1.layer.1.layer_norm.weight": "model-00001-of-00004.safetensors",
25
+ "encoder.block.10.layer.0.SelfAttention.k.weight": "model-00002-of-00004.safetensors",
26
+ "encoder.block.10.layer.0.SelfAttention.o.weight": "model-00002-of-00004.safetensors",
27
+ "encoder.block.10.layer.0.SelfAttention.q.weight": "model-00002-of-00004.safetensors",
28
+ "encoder.block.10.layer.0.SelfAttention.v.weight": "model-00002-of-00004.safetensors",
29
+ "encoder.block.10.layer.0.layer_norm.weight": "model-00002-of-00004.safetensors",
30
+ "encoder.block.10.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00004.safetensors",
31
+ "encoder.block.10.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00004.safetensors",
32
+ "encoder.block.10.layer.1.DenseReluDense.wo.weight": "model-00002-of-00004.safetensors",
33
+ "encoder.block.10.layer.1.layer_norm.weight": "model-00002-of-00004.safetensors",
34
+ "encoder.block.11.layer.0.SelfAttention.k.weight": "model-00002-of-00004.safetensors",
35
+ "encoder.block.11.layer.0.SelfAttention.o.weight": "model-00002-of-00004.safetensors",
36
+ "encoder.block.11.layer.0.SelfAttention.q.weight": "model-00002-of-00004.safetensors",
37
+ "encoder.block.11.layer.0.SelfAttention.v.weight": "model-00002-of-00004.safetensors",
38
+ "encoder.block.11.layer.0.layer_norm.weight": "model-00002-of-00004.safetensors",
39
+ "encoder.block.11.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00004.safetensors",
40
+ "encoder.block.11.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00004.safetensors",
41
+ "encoder.block.11.layer.1.DenseReluDense.wo.weight": "model-00002-of-00004.safetensors",
42
+ "encoder.block.11.layer.1.layer_norm.weight": "model-00002-of-00004.safetensors",
43
+ "encoder.block.12.layer.0.SelfAttention.k.weight": "model-00002-of-00004.safetensors",
44
+ "encoder.block.12.layer.0.SelfAttention.o.weight": "model-00003-of-00004.safetensors",
45
+ "encoder.block.12.layer.0.SelfAttention.q.weight": "model-00002-of-00004.safetensors",
46
+ "encoder.block.12.layer.0.SelfAttention.v.weight": "model-00002-of-00004.safetensors",
47
+ "encoder.block.12.layer.0.layer_norm.weight": "model-00003-of-00004.safetensors",
48
+ "encoder.block.12.layer.1.DenseReluDense.wi_0.weight": "model-00003-of-00004.safetensors",
49
+ "encoder.block.12.layer.1.DenseReluDense.wi_1.weight": "model-00003-of-00004.safetensors",
50
+ "encoder.block.12.layer.1.DenseReluDense.wo.weight": "model-00003-of-00004.safetensors",
51
+ "encoder.block.12.layer.1.layer_norm.weight": "model-00003-of-00004.safetensors",
52
+ "encoder.block.13.layer.0.SelfAttention.k.weight": "model-00003-of-00004.safetensors",
53
+ "encoder.block.13.layer.0.SelfAttention.o.weight": "model-00003-of-00004.safetensors",
54
+ "encoder.block.13.layer.0.SelfAttention.q.weight": "model-00003-of-00004.safetensors",
55
+ "encoder.block.13.layer.0.SelfAttention.v.weight": "model-00003-of-00004.safetensors",
56
+ "encoder.block.13.layer.0.layer_norm.weight": "model-00003-of-00004.safetensors",
57
+ "encoder.block.13.layer.1.DenseReluDense.wi_0.weight": "model-00003-of-00004.safetensors",
58
+ "encoder.block.13.layer.1.DenseReluDense.wi_1.weight": "model-00003-of-00004.safetensors",
59
+ "encoder.block.13.layer.1.DenseReluDense.wo.weight": "model-00003-of-00004.safetensors",
60
+ "encoder.block.13.layer.1.layer_norm.weight": "model-00003-of-00004.safetensors",
61
+ "encoder.block.14.layer.0.SelfAttention.k.weight": "model-00003-of-00004.safetensors",
62
+ "encoder.block.14.layer.0.SelfAttention.o.weight": "model-00003-of-00004.safetensors",
63
+ "encoder.block.14.layer.0.SelfAttention.q.weight": "model-00003-of-00004.safetensors",
64
+ "encoder.block.14.layer.0.SelfAttention.v.weight": "model-00003-of-00004.safetensors",
65
+ "encoder.block.14.layer.0.layer_norm.weight": "model-00003-of-00004.safetensors",
66
+ "encoder.block.14.layer.1.DenseReluDense.wi_0.weight": "model-00003-of-00004.safetensors",
67
+ "encoder.block.14.layer.1.DenseReluDense.wi_1.weight": "model-00003-of-00004.safetensors",
68
+ "encoder.block.14.layer.1.DenseReluDense.wo.weight": "model-00003-of-00004.safetensors",
69
+ "encoder.block.14.layer.1.layer_norm.weight": "model-00003-of-00004.safetensors",
70
+ "encoder.block.15.layer.0.SelfAttention.k.weight": "model-00003-of-00004.safetensors",
71
+ "encoder.block.15.layer.0.SelfAttention.o.weight": "model-00003-of-00004.safetensors",
72
+ "encoder.block.15.layer.0.SelfAttention.q.weight": "model-00003-of-00004.safetensors",
73
+ "encoder.block.15.layer.0.SelfAttention.v.weight": "model-00003-of-00004.safetensors",
74
+ "encoder.block.15.layer.0.layer_norm.weight": "model-00003-of-00004.safetensors",
75
+ "encoder.block.15.layer.1.DenseReluDense.wi_0.weight": "model-00003-of-00004.safetensors",
76
+ "encoder.block.15.layer.1.DenseReluDense.wi_1.weight": "model-00003-of-00004.safetensors",
77
+ "encoder.block.15.layer.1.DenseReluDense.wo.weight": "model-00003-of-00004.safetensors",
78
+ "encoder.block.15.layer.1.layer_norm.weight": "model-00003-of-00004.safetensors",
79
+ "encoder.block.16.layer.0.SelfAttention.k.weight": "model-00003-of-00004.safetensors",
80
+ "encoder.block.16.layer.0.SelfAttention.o.weight": "model-00003-of-00004.safetensors",
81
+ "encoder.block.16.layer.0.SelfAttention.q.weight": "model-00003-of-00004.safetensors",
82
+ "encoder.block.16.layer.0.SelfAttention.v.weight": "model-00003-of-00004.safetensors",
83
+ "encoder.block.16.layer.0.layer_norm.weight": "model-00003-of-00004.safetensors",
84
+ "encoder.block.16.layer.1.DenseReluDense.wi_0.weight": "model-00003-of-00004.safetensors",
85
+ "encoder.block.16.layer.1.DenseReluDense.wi_1.weight": "model-00003-of-00004.safetensors",
86
+ "encoder.block.16.layer.1.DenseReluDense.wo.weight": "model-00003-of-00004.safetensors",
87
+ "encoder.block.16.layer.1.layer_norm.weight": "model-00003-of-00004.safetensors",
88
+ "encoder.block.17.layer.0.SelfAttention.k.weight": "model-00003-of-00004.safetensors",
89
+ "encoder.block.17.layer.0.SelfAttention.o.weight": "model-00003-of-00004.safetensors",
90
+ "encoder.block.17.layer.0.SelfAttention.q.weight": "model-00003-of-00004.safetensors",
91
+ "encoder.block.17.layer.0.SelfAttention.v.weight": "model-00003-of-00004.safetensors",
92
+ "encoder.block.17.layer.0.layer_norm.weight": "model-00003-of-00004.safetensors",
93
+ "encoder.block.17.layer.1.DenseReluDense.wi_0.weight": "model-00003-of-00004.safetensors",
94
+ "encoder.block.17.layer.1.DenseReluDense.wi_1.weight": "model-00003-of-00004.safetensors",
95
+ "encoder.block.17.layer.1.DenseReluDense.wo.weight": "model-00003-of-00004.safetensors",
96
+ "encoder.block.17.layer.1.layer_norm.weight": "model-00003-of-00004.safetensors",
97
+ "encoder.block.18.layer.0.SelfAttention.k.weight": "model-00003-of-00004.safetensors",
98
+ "encoder.block.18.layer.0.SelfAttention.o.weight": "model-00003-of-00004.safetensors",
99
+ "encoder.block.18.layer.0.SelfAttention.q.weight": "model-00003-of-00004.safetensors",
100
+ "encoder.block.18.layer.0.SelfAttention.v.weight": "model-00003-of-00004.safetensors",
101
+ "encoder.block.18.layer.0.layer_norm.weight": "model-00003-of-00004.safetensors",
102
+ "encoder.block.18.layer.1.DenseReluDense.wi_0.weight": "model-00003-of-00004.safetensors",
103
+ "encoder.block.18.layer.1.DenseReluDense.wi_1.weight": "model-00004-of-00004.safetensors",
104
+ "encoder.block.18.layer.1.DenseReluDense.wo.weight": "model-00004-of-00004.safetensors",
105
+ "encoder.block.18.layer.1.layer_norm.weight": "model-00004-of-00004.safetensors",
106
+ "encoder.block.19.layer.0.SelfAttention.k.weight": "model-00004-of-00004.safetensors",
107
+ "encoder.block.19.layer.0.SelfAttention.o.weight": "model-00004-of-00004.safetensors",
108
+ "encoder.block.19.layer.0.SelfAttention.q.weight": "model-00004-of-00004.safetensors",
109
+ "encoder.block.19.layer.0.SelfAttention.v.weight": "model-00004-of-00004.safetensors",
110
+ "encoder.block.19.layer.0.layer_norm.weight": "model-00004-of-00004.safetensors",
111
+ "encoder.block.19.layer.1.DenseReluDense.wi_0.weight": "model-00004-of-00004.safetensors",
112
+ "encoder.block.19.layer.1.DenseReluDense.wi_1.weight": "model-00004-of-00004.safetensors",
113
+ "encoder.block.19.layer.1.DenseReluDense.wo.weight": "model-00004-of-00004.safetensors",
114
+ "encoder.block.19.layer.1.layer_norm.weight": "model-00004-of-00004.safetensors",
115
+ "encoder.block.2.layer.0.SelfAttention.k.weight": "model-00001-of-00004.safetensors",
116
+ "encoder.block.2.layer.0.SelfAttention.o.weight": "model-00001-of-00004.safetensors",
117
+ "encoder.block.2.layer.0.SelfAttention.q.weight": "model-00001-of-00004.safetensors",
118
+ "encoder.block.2.layer.0.SelfAttention.v.weight": "model-00001-of-00004.safetensors",
119
+ "encoder.block.2.layer.0.layer_norm.weight": "model-00001-of-00004.safetensors",
120
+ "encoder.block.2.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00004.safetensors",
121
+ "encoder.block.2.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00004.safetensors",
122
+ "encoder.block.2.layer.1.DenseReluDense.wo.weight": "model-00001-of-00004.safetensors",
123
+ "encoder.block.2.layer.1.layer_norm.weight": "model-00001-of-00004.safetensors",
124
+ "encoder.block.20.layer.0.SelfAttention.k.weight": "model-00004-of-00004.safetensors",
125
+ "encoder.block.20.layer.0.SelfAttention.o.weight": "model-00004-of-00004.safetensors",
126
+ "encoder.block.20.layer.0.SelfAttention.q.weight": "model-00004-of-00004.safetensors",
127
+ "encoder.block.20.layer.0.SelfAttention.v.weight": "model-00004-of-00004.safetensors",
128
+ "encoder.block.20.layer.0.layer_norm.weight": "model-00004-of-00004.safetensors",
129
+ "encoder.block.20.layer.1.DenseReluDense.wi_0.weight": "model-00004-of-00004.safetensors",
130
+ "encoder.block.20.layer.1.DenseReluDense.wi_1.weight": "model-00004-of-00004.safetensors",
131
+ "encoder.block.20.layer.1.DenseReluDense.wo.weight": "model-00004-of-00004.safetensors",
132
+ "encoder.block.20.layer.1.layer_norm.weight": "model-00004-of-00004.safetensors",
133
+ "encoder.block.21.layer.0.SelfAttention.k.weight": "model-00004-of-00004.safetensors",
134
+ "encoder.block.21.layer.0.SelfAttention.o.weight": "model-00004-of-00004.safetensors",
135
+ "encoder.block.21.layer.0.SelfAttention.q.weight": "model-00004-of-00004.safetensors",
136
+ "encoder.block.21.layer.0.SelfAttention.v.weight": "model-00004-of-00004.safetensors",
137
+ "encoder.block.21.layer.0.layer_norm.weight": "model-00004-of-00004.safetensors",
138
+ "encoder.block.21.layer.1.DenseReluDense.wi_0.weight": "model-00004-of-00004.safetensors",
139
+ "encoder.block.21.layer.1.DenseReluDense.wi_1.weight": "model-00004-of-00004.safetensors",
140
+ "encoder.block.21.layer.1.DenseReluDense.wo.weight": "model-00004-of-00004.safetensors",
141
+ "encoder.block.21.layer.1.layer_norm.weight": "model-00004-of-00004.safetensors",
142
+ "encoder.block.22.layer.0.SelfAttention.k.weight": "model-00004-of-00004.safetensors",
143
+ "encoder.block.22.layer.0.SelfAttention.o.weight": "model-00004-of-00004.safetensors",
144
+ "encoder.block.22.layer.0.SelfAttention.q.weight": "model-00004-of-00004.safetensors",
145
+ "encoder.block.22.layer.0.SelfAttention.v.weight": "model-00004-of-00004.safetensors",
146
+ "encoder.block.22.layer.0.layer_norm.weight": "model-00004-of-00004.safetensors",
147
+ "encoder.block.22.layer.1.DenseReluDense.wi_0.weight": "model-00004-of-00004.safetensors",
148
+ "encoder.block.22.layer.1.DenseReluDense.wi_1.weight": "model-00004-of-00004.safetensors",
149
+ "encoder.block.22.layer.1.DenseReluDense.wo.weight": "model-00004-of-00004.safetensors",
150
+ "encoder.block.22.layer.1.layer_norm.weight": "model-00004-of-00004.safetensors",
151
+ "encoder.block.23.layer.0.SelfAttention.k.weight": "model-00004-of-00004.safetensors",
152
+ "encoder.block.23.layer.0.SelfAttention.o.weight": "model-00004-of-00004.safetensors",
153
+ "encoder.block.23.layer.0.SelfAttention.q.weight": "model-00004-of-00004.safetensors",
154
+ "encoder.block.23.layer.0.SelfAttention.v.weight": "model-00004-of-00004.safetensors",
155
+ "encoder.block.23.layer.0.layer_norm.weight": "model-00004-of-00004.safetensors",
156
+ "encoder.block.23.layer.1.DenseReluDense.wi_0.weight": "model-00004-of-00004.safetensors",
157
+ "encoder.block.23.layer.1.DenseReluDense.wi_1.weight": "model-00004-of-00004.safetensors",
158
+ "encoder.block.23.layer.1.DenseReluDense.wo.weight": "model-00004-of-00004.safetensors",
159
+ "encoder.block.23.layer.1.layer_norm.weight": "model-00004-of-00004.safetensors",
160
+ "encoder.block.3.layer.0.SelfAttention.k.weight": "model-00001-of-00004.safetensors",
161
+ "encoder.block.3.layer.0.SelfAttention.o.weight": "model-00001-of-00004.safetensors",
162
+ "encoder.block.3.layer.0.SelfAttention.q.weight": "model-00001-of-00004.safetensors",
163
+ "encoder.block.3.layer.0.SelfAttention.v.weight": "model-00001-of-00004.safetensors",
164
+ "encoder.block.3.layer.0.layer_norm.weight": "model-00001-of-00004.safetensors",
165
+ "encoder.block.3.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00004.safetensors",
166
+ "encoder.block.3.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00004.safetensors",
167
+ "encoder.block.3.layer.1.DenseReluDense.wo.weight": "model-00001-of-00004.safetensors",
168
+ "encoder.block.3.layer.1.layer_norm.weight": "model-00001-of-00004.safetensors",
169
+ "encoder.block.4.layer.0.SelfAttention.k.weight": "model-00001-of-00004.safetensors",
170
+ "encoder.block.4.layer.0.SelfAttention.o.weight": "model-00001-of-00004.safetensors",
171
+ "encoder.block.4.layer.0.SelfAttention.q.weight": "model-00001-of-00004.safetensors",
172
+ "encoder.block.4.layer.0.SelfAttention.v.weight": "model-00001-of-00004.safetensors",
173
+ "encoder.block.4.layer.0.layer_norm.weight": "model-00001-of-00004.safetensors",
174
+ "encoder.block.4.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00004.safetensors",
175
+ "encoder.block.4.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00004.safetensors",
176
+ "encoder.block.4.layer.1.DenseReluDense.wo.weight": "model-00001-of-00004.safetensors",
177
+ "encoder.block.4.layer.1.layer_norm.weight": "model-00001-of-00004.safetensors",
178
+ "encoder.block.5.layer.0.SelfAttention.k.weight": "model-00001-of-00004.safetensors",
179
+ "encoder.block.5.layer.0.SelfAttention.o.weight": "model-00001-of-00004.safetensors",
180
+ "encoder.block.5.layer.0.SelfAttention.q.weight": "model-00001-of-00004.safetensors",
181
+ "encoder.block.5.layer.0.SelfAttention.v.weight": "model-00001-of-00004.safetensors",
182
+ "encoder.block.5.layer.0.layer_norm.weight": "model-00001-of-00004.safetensors",
183
+ "encoder.block.5.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00004.safetensors",
184
+ "encoder.block.5.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00004.safetensors",
185
+ "encoder.block.5.layer.1.DenseReluDense.wo.weight": "model-00002-of-00004.safetensors",
186
+ "encoder.block.5.layer.1.layer_norm.weight": "model-00002-of-00004.safetensors",
187
+ "encoder.block.6.layer.0.SelfAttention.k.weight": "model-00002-of-00004.safetensors",
188
+ "encoder.block.6.layer.0.SelfAttention.o.weight": "model-00002-of-00004.safetensors",
189
+ "encoder.block.6.layer.0.SelfAttention.q.weight": "model-00002-of-00004.safetensors",
190
+ "encoder.block.6.layer.0.SelfAttention.v.weight": "model-00002-of-00004.safetensors",
191
+ "encoder.block.6.layer.0.layer_norm.weight": "model-00002-of-00004.safetensors",
192
+ "encoder.block.6.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00004.safetensors",
193
+ "encoder.block.6.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00004.safetensors",
194
+ "encoder.block.6.layer.1.DenseReluDense.wo.weight": "model-00002-of-00004.safetensors",
195
+ "encoder.block.6.layer.1.layer_norm.weight": "model-00002-of-00004.safetensors",
196
+ "encoder.block.7.layer.0.SelfAttention.k.weight": "model-00002-of-00004.safetensors",
197
+ "encoder.block.7.layer.0.SelfAttention.o.weight": "model-00002-of-00004.safetensors",
198
+ "encoder.block.7.layer.0.SelfAttention.q.weight": "model-00002-of-00004.safetensors",
199
+ "encoder.block.7.layer.0.SelfAttention.v.weight": "model-00002-of-00004.safetensors",
200
+ "encoder.block.7.layer.0.layer_norm.weight": "model-00002-of-00004.safetensors",
201
+ "encoder.block.7.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00004.safetensors",
202
+ "encoder.block.7.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00004.safetensors",
203
+ "encoder.block.7.layer.1.DenseReluDense.wo.weight": "model-00002-of-00004.safetensors",
204
+ "encoder.block.7.layer.1.layer_norm.weight": "model-00002-of-00004.safetensors",
205
+ "encoder.block.8.layer.0.SelfAttention.k.weight": "model-00002-of-00004.safetensors",
206
+ "encoder.block.8.layer.0.SelfAttention.o.weight": "model-00002-of-00004.safetensors",
207
+ "encoder.block.8.layer.0.SelfAttention.q.weight": "model-00002-of-00004.safetensors",
208
+ "encoder.block.8.layer.0.SelfAttention.v.weight": "model-00002-of-00004.safetensors",
209
+ "encoder.block.8.layer.0.layer_norm.weight": "model-00002-of-00004.safetensors",
210
+ "encoder.block.8.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00004.safetensors",
211
+ "encoder.block.8.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00004.safetensors",
212
+ "encoder.block.8.layer.1.DenseReluDense.wo.weight": "model-00002-of-00004.safetensors",
213
+ "encoder.block.8.layer.1.layer_norm.weight": "model-00002-of-00004.safetensors",
214
+ "encoder.block.9.layer.0.SelfAttention.k.weight": "model-00002-of-00004.safetensors",
215
+ "encoder.block.9.layer.0.SelfAttention.o.weight": "model-00002-of-00004.safetensors",
216
+ "encoder.block.9.layer.0.SelfAttention.q.weight": "model-00002-of-00004.safetensors",
217
+ "encoder.block.9.layer.0.SelfAttention.v.weight": "model-00002-of-00004.safetensors",
218
+ "encoder.block.9.layer.0.layer_norm.weight": "model-00002-of-00004.safetensors",
219
+ "encoder.block.9.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00004.safetensors",
220
+ "encoder.block.9.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00004.safetensors",
221
+ "encoder.block.9.layer.1.DenseReluDense.wo.weight": "model-00002-of-00004.safetensors",
222
+ "encoder.block.9.layer.1.layer_norm.weight": "model-00002-of-00004.safetensors",
223
+ "encoder.final_layer_norm.weight": "model-00004-of-00004.safetensors",
224
+ "shared.weight": "model-00001-of-00004.safetensors"
225
  }
226
  }
tokenizer/tokenizer_config.json CHANGED
@@ -932,7 +932,7 @@
932
  "eos_token": "</s>",
933
  "extra_ids": 100,
934
  "legacy": true,
935
- "model_max_length": 256,
936
  "pad_token": "<pad>",
937
  "sp_model_kwargs": {},
938
  "tokenizer_class": "T5Tokenizer",
 
932
  "eos_token": "</s>",
933
  "extra_ids": 100,
934
  "legacy": true,
935
+ "model_max_length": 512,
936
  "pad_token": "<pad>",
937
  "sp_model_kwargs": {},
938
  "tokenizer_class": "T5Tokenizer",