lucyknada commited on
Commit
12e7b0b
1 Parent(s): 4ccb1c3

Upload folder using huggingface_hub

Browse files
README.md CHANGED
@@ -1,3 +1,124 @@
1
  ---
2
- license: cc-by-4.0
 
 
 
 
 
 
 
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ language:
3
+ - eng
4
+ tags:
5
+ - sft
6
+ - Yi-34B-200K
7
+ license:
8
+ - mit
9
+ datasets:
10
+ - LDJnr/Capybara
11
+ - LDJnr/LessWrong-Amplify-Instruct
12
+ - LDJnr/Pure-Dove
13
+ - LDJnr/Verified-Camel
14
  ---
15
+
16
+ ## **Nous-Capybara-34B V1.9**
17
+
18
+ **This is trained on the Yi-34B model with 200K context length, for 3 epochs on the Capybara dataset!**
19
+
20
+ **First 34B Nous model and first 200K context length Nous model!**
21
+
22
+ The Capybara series is the first Nous collection of models made by fine-tuning mostly on data created by Nous in-house.
23
+
24
+ We leverage our novel data synthesis technique called Amplify-instruct (Paper coming soon), the seed distribution and synthesis method are comprised of a synergistic combination of top performing existing data synthesis techniques and distributions used for SOTA models such as Airoboros, Evol-Instruct(WizardLM), Orca, Vicuna, Know_Logic, Lamini, FLASK and others, all into one lean holistically formed methodology for the dataset and model. The seed instructions used for the start of synthesized conversations are largely based on highly regarded datasets like Airoboros, Know logic, EverythingLM, GPTeacher and even entirely new seed instructions derived from posts on the website LessWrong, as well as being supplemented with certain in-house multi-turn datasets like Dove(A successor to Puffin).
25
+
26
+ While performing great in it's current state, the current dataset used for fine-tuning is entirely contained within 20K training examples, this is 10 times smaller than many similar performing current models, this is signficant when it comes to scaling implications for our next generation of models once we scale our novel syntheiss methods to significantly more examples.
27
+
28
+ ## Process of creation and special thank yous!
29
+
30
+ This model was fine-tuned by Nous Research as part of the Capybara/Amplify-Instruct project led by Luigi D.(LDJ) (Paper coming soon), as well as significant dataset formation contributions by J-Supha and general compute and experimentation management by Jeffrey Q. during ablations.
31
+
32
+ Special thank you to **A16Z** for sponsoring our training, as well as **Yield Protocol** for their support in financially sponsoring resources during the R&D of this project.
33
+
34
+ ## Thank you to those of you that have indirectly contributed!
35
+
36
+ While most of the tokens within Capybara are newly synthsized and part of datasets like Puffin/Dove, we would like to credit the single-turn datasets we leveraged as seeds that are used to generate the multi-turn data as part of the Amplify-Instruct synthesis.
37
+
38
+ The datasets shown in green below are datasets that we sampled from to curate seeds that are used during Amplify-Instruct synthesis for this project.
39
+
40
+ Datasets in Blue are in-house curations that previously existed prior to Capybara.
41
+
42
+ ![Capybara](https://i.imgur.com/yB58OoD.jpeg)
43
+
44
+
45
+ ## Prompt Format
46
+
47
+ The reccomended model usage is:
48
+
49
+
50
+ Prefix: ``USER:``
51
+
52
+ Suffix: ``ASSISTANT:``
53
+
54
+ Stop token: ``</s>``
55
+
56
+
57
+ ## Mutli-Modality!
58
+
59
+ - We currently have a Multi-modal model based on Capybara V1.9!
60
+ https://huggingface.co/NousResearch/Obsidian-3B-V0.5
61
+ it is currently only available as a 3B sized model but larger versions coming!
62
+
63
+
64
+ ## Notable Features:
65
+
66
+ - Uses Yi-34B model as the base which is trained for 200K context length!
67
+
68
+ - Over 60% of the dataset is comprised of multi-turn conversations.(Most models are still only trained for single-turn conversations and no back and forths!)
69
+
70
+ - Over 1,000 tokens average per conversation example! (Most models are trained on conversation data that is less than 300 tokens per example.)
71
+
72
+ - Able to effectively do complex summaries of advanced topics and studies. (trained on hundreds of advanced difficult summary tasks developed in-house)
73
+
74
+ - Ability to recall information upto late 2022 without internet.
75
+
76
+ - Includes a portion of conversational data synthesized from less wrong posts, discussing very in-depth details and philosophies about the nature of reality, reasoning, rationality, self-improvement and related concepts.
77
+
78
+ ## Example Outputs from Capybara V1.9 7B version! (examples from 34B coming soon):
79
+
80
+ ![Capybara](https://img001.prntscr.com/file/img001/T9yYxR1xQSaK_UGdy3t2Cw.png)
81
+
82
+ ![Capybara](https://img001.prntscr.com/file/img001/DQXqmKbsQQOIcgny1eoGNA.png)
83
+
84
+ ![Capybara](https://img001.prntscr.com/file/img001/85X3L9ZxTsOKo3fUQ7GRVA.png)
85
+
86
+ ## Benchmarks! (Coming soon!)
87
+
88
+
89
+ ## Future model sizes
90
+
91
+ Capybara V1.9 now currently has a 3B, 7B and 34B size, and we plan to eventually have a 13B and 70B version in the future, as well as a potential 1B version based on phi-1.5 or Tiny Llama.
92
+
93
+ ## How you can help!
94
+
95
+ In the near future we plan on leveraging the help of domain specific expert volunteers to eliminate any mathematically/verifiably incorrect answers from our training curations.
96
+
97
+ If you have at-least a bachelors in mathematics, physics, biology or chemistry and would like to volunteer even just 30 minutes of your expertise time, please contact LDJ on discord!
98
+
99
+ ## Dataset contamination.
100
+
101
+ We have checked the capybara dataset for contamination for several of the most popular datasets and can confirm that there is no contaminaton found.
102
+
103
+ We leveraged minhash to check for 100%, 99%, 98% and 97% similarity matches between our data and the questions and answers in benchmarks, we found no exact matches, nor did we find any matches down to the 97% similarity level.
104
+
105
+ The following are benchmarks we checked for contamination against our dataset:
106
+
107
+ - HumanEval
108
+
109
+ - AGIEval
110
+
111
+ - TruthfulQA
112
+
113
+ - MMLU
114
+
115
+ - GPT4All
116
+
117
+ ```
118
+ @article{daniele2023amplify-instruct,
119
+ title={Amplify-Instruct: Synthetically Generated Diverse Multi-turn Conversations for Effecient LLM Training.},
120
+ author={Daniele, Luigi and Suphavadeeprasit},
121
+ journal={arXiv preprint arXiv:(comming soon)},
122
+ year={2023}
123
+ }
124
+ ```
config.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "larryvrh/Yi-34B-200K-Llamafied",
3
+ "architectures": [
4
+ "LlamaForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "bos_token_id": 1,
8
+ "eos_token_id": 2,
9
+ "hidden_act": "silu",
10
+ "hidden_size": 7168,
11
+ "initializer_range": 0.02,
12
+ "intermediate_size": 20480,
13
+ "max_position_embeddings": 200000,
14
+ "model_type": "llama",
15
+ "num_attention_heads": 56,
16
+ "num_hidden_layers": 60,
17
+ "num_key_value_heads": 8,
18
+ "pad_token_id": 0,
19
+ "pretraining_tp": 1,
20
+ "rms_norm_eps": 1e-05,
21
+ "rope_scaling": null,
22
+ "rope_theta": 5000000.0,
23
+ "tie_word_embeddings": false,
24
+ "torch_dtype": "bfloat16",
25
+ "transformers_version": "4.34.1",
26
+ "use_cache": true,
27
+ "vocab_size": 64000
28
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.34.1"
7
+ }
output-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6756996ec4830f69fae3b25adaf0887126856c73278851720c2887ac3f4afd52
3
+ size 8532901616
output-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49f7eacf5fc1892cc0797b4653a7413f9ccf6ccaa6f51480328815b43f882d8c
3
+ size 5317444376
pytorch_model.bin.index.json ADDED
@@ -0,0 +1,550 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 68777834496
4
+ },
5
+ "weight_map": {
6
+ "lm_head.weight": "pytorch_model-00007-of-00007.bin",
7
+ "model.embed_tokens.weight": "pytorch_model-00001-of-00007.bin",
8
+ "model.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00007.bin",
9
+ "model.layers.0.mlp.down_proj.weight": "pytorch_model-00001-of-00007.bin",
10
+ "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00001-of-00007.bin",
11
+ "model.layers.0.mlp.up_proj.weight": "pytorch_model-00001-of-00007.bin",
12
+ "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00007.bin",
13
+ "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
14
+ "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00007.bin",
15
+ "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
16
+ "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
17
+ "model.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00007.bin",
18
+ "model.layers.1.mlp.down_proj.weight": "pytorch_model-00001-of-00007.bin",
19
+ "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00001-of-00007.bin",
20
+ "model.layers.1.mlp.up_proj.weight": "pytorch_model-00001-of-00007.bin",
21
+ "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00007.bin",
22
+ "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
23
+ "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00001-of-00007.bin",
24
+ "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
25
+ "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
26
+ "model.layers.10.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
27
+ "model.layers.10.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
28
+ "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
29
+ "model.layers.10.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
30
+ "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
31
+ "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00002-of-00007.bin",
32
+ "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
33
+ "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00002-of-00007.bin",
34
+ "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00002-of-00007.bin",
35
+ "model.layers.11.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
36
+ "model.layers.11.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
37
+ "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
38
+ "model.layers.11.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
39
+ "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
40
+ "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00002-of-00007.bin",
41
+ "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
42
+ "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00002-of-00007.bin",
43
+ "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00002-of-00007.bin",
44
+ "model.layers.12.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
45
+ "model.layers.12.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
46
+ "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
47
+ "model.layers.12.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
48
+ "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
49
+ "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00002-of-00007.bin",
50
+ "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
51
+ "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00002-of-00007.bin",
52
+ "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00002-of-00007.bin",
53
+ "model.layers.13.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
54
+ "model.layers.13.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
55
+ "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
56
+ "model.layers.13.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
57
+ "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
58
+ "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00002-of-00007.bin",
59
+ "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
60
+ "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00002-of-00007.bin",
61
+ "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00002-of-00007.bin",
62
+ "model.layers.14.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
63
+ "model.layers.14.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
64
+ "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
65
+ "model.layers.14.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
66
+ "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
67
+ "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00002-of-00007.bin",
68
+ "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
69
+ "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00002-of-00007.bin",
70
+ "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00002-of-00007.bin",
71
+ "model.layers.15.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
72
+ "model.layers.15.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
73
+ "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
74
+ "model.layers.15.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
75
+ "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
76
+ "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00002-of-00007.bin",
77
+ "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
78
+ "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00002-of-00007.bin",
79
+ "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00002-of-00007.bin",
80
+ "model.layers.16.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
81
+ "model.layers.16.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
82
+ "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
83
+ "model.layers.16.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
84
+ "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
85
+ "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00002-of-00007.bin",
86
+ "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
87
+ "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00002-of-00007.bin",
88
+ "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00002-of-00007.bin",
89
+ "model.layers.17.input_layernorm.weight": "pytorch_model-00003-of-00007.bin",
90
+ "model.layers.17.mlp.down_proj.weight": "pytorch_model-00003-of-00007.bin",
91
+ "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
92
+ "model.layers.17.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
93
+ "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00003-of-00007.bin",
94
+ "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
95
+ "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
96
+ "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
97
+ "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
98
+ "model.layers.18.input_layernorm.weight": "pytorch_model-00003-of-00007.bin",
99
+ "model.layers.18.mlp.down_proj.weight": "pytorch_model-00003-of-00007.bin",
100
+ "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
101
+ "model.layers.18.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
102
+ "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00003-of-00007.bin",
103
+ "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
104
+ "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
105
+ "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
106
+ "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
107
+ "model.layers.19.input_layernorm.weight": "pytorch_model-00003-of-00007.bin",
108
+ "model.layers.19.mlp.down_proj.weight": "pytorch_model-00003-of-00007.bin",
109
+ "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
110
+ "model.layers.19.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
111
+ "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00003-of-00007.bin",
112
+ "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
113
+ "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
114
+ "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
115
+ "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
116
+ "model.layers.2.input_layernorm.weight": "pytorch_model-00001-of-00007.bin",
117
+ "model.layers.2.mlp.down_proj.weight": "pytorch_model-00001-of-00007.bin",
118
+ "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00001-of-00007.bin",
119
+ "model.layers.2.mlp.up_proj.weight": "pytorch_model-00001-of-00007.bin",
120
+ "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00007.bin",
121
+ "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
122
+ "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00001-of-00007.bin",
123
+ "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
124
+ "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
125
+ "model.layers.20.input_layernorm.weight": "pytorch_model-00003-of-00007.bin",
126
+ "model.layers.20.mlp.down_proj.weight": "pytorch_model-00003-of-00007.bin",
127
+ "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
128
+ "model.layers.20.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
129
+ "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00003-of-00007.bin",
130
+ "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
131
+ "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
132
+ "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
133
+ "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
134
+ "model.layers.21.input_layernorm.weight": "pytorch_model-00003-of-00007.bin",
135
+ "model.layers.21.mlp.down_proj.weight": "pytorch_model-00003-of-00007.bin",
136
+ "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
137
+ "model.layers.21.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
138
+ "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00003-of-00007.bin",
139
+ "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
140
+ "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
141
+ "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
142
+ "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
143
+ "model.layers.22.input_layernorm.weight": "pytorch_model-00003-of-00007.bin",
144
+ "model.layers.22.mlp.down_proj.weight": "pytorch_model-00003-of-00007.bin",
145
+ "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
146
+ "model.layers.22.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
147
+ "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00003-of-00007.bin",
148
+ "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
149
+ "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
150
+ "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
151
+ "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
152
+ "model.layers.23.input_layernorm.weight": "pytorch_model-00003-of-00007.bin",
153
+ "model.layers.23.mlp.down_proj.weight": "pytorch_model-00003-of-00007.bin",
154
+ "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
155
+ "model.layers.23.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
156
+ "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00003-of-00007.bin",
157
+ "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
158
+ "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
159
+ "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
160
+ "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
161
+ "model.layers.24.input_layernorm.weight": "pytorch_model-00003-of-00007.bin",
162
+ "model.layers.24.mlp.down_proj.weight": "pytorch_model-00003-of-00007.bin",
163
+ "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
164
+ "model.layers.24.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
165
+ "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00003-of-00007.bin",
166
+ "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
167
+ "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
168
+ "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
169
+ "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
170
+ "model.layers.25.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
171
+ "model.layers.25.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
172
+ "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
173
+ "model.layers.25.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
174
+ "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
175
+ "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
176
+ "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
177
+ "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
178
+ "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
179
+ "model.layers.26.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
180
+ "model.layers.26.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
181
+ "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
182
+ "model.layers.26.mlp.up_proj.weight": "pytorch_model-00004-of-00007.bin",
183
+ "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
184
+ "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
185
+ "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
186
+ "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
187
+ "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
188
+ "model.layers.27.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
189
+ "model.layers.27.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
190
+ "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
191
+ "model.layers.27.mlp.up_proj.weight": "pytorch_model-00004-of-00007.bin",
192
+ "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
193
+ "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
194
+ "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
195
+ "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
196
+ "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
197
+ "model.layers.28.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
198
+ "model.layers.28.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
199
+ "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
200
+ "model.layers.28.mlp.up_proj.weight": "pytorch_model-00004-of-00007.bin",
201
+ "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
202
+ "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
203
+ "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
204
+ "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
205
+ "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
206
+ "model.layers.29.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
207
+ "model.layers.29.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
208
+ "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
209
+ "model.layers.29.mlp.up_proj.weight": "pytorch_model-00004-of-00007.bin",
210
+ "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
211
+ "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
212
+ "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
213
+ "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
214
+ "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
215
+ "model.layers.3.input_layernorm.weight": "pytorch_model-00001-of-00007.bin",
216
+ "model.layers.3.mlp.down_proj.weight": "pytorch_model-00001-of-00007.bin",
217
+ "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00001-of-00007.bin",
218
+ "model.layers.3.mlp.up_proj.weight": "pytorch_model-00001-of-00007.bin",
219
+ "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00007.bin",
220
+ "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
221
+ "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00001-of-00007.bin",
222
+ "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
223
+ "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
224
+ "model.layers.30.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
225
+ "model.layers.30.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
226
+ "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
227
+ "model.layers.30.mlp.up_proj.weight": "pytorch_model-00004-of-00007.bin",
228
+ "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
229
+ "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
230
+ "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
231
+ "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
232
+ "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
233
+ "model.layers.31.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
234
+ "model.layers.31.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
235
+ "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
236
+ "model.layers.31.mlp.up_proj.weight": "pytorch_model-00004-of-00007.bin",
237
+ "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
238
+ "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
239
+ "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
240
+ "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
241
+ "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
242
+ "model.layers.32.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
243
+ "model.layers.32.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
244
+ "model.layers.32.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
245
+ "model.layers.32.mlp.up_proj.weight": "pytorch_model-00004-of-00007.bin",
246
+ "model.layers.32.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
247
+ "model.layers.32.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
248
+ "model.layers.32.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
249
+ "model.layers.32.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
250
+ "model.layers.32.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
251
+ "model.layers.33.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
252
+ "model.layers.33.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
253
+ "model.layers.33.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
254
+ "model.layers.33.mlp.up_proj.weight": "pytorch_model-00004-of-00007.bin",
255
+ "model.layers.33.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
256
+ "model.layers.33.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
257
+ "model.layers.33.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
258
+ "model.layers.33.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
259
+ "model.layers.33.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
260
+ "model.layers.34.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
261
+ "model.layers.34.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
262
+ "model.layers.34.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
263
+ "model.layers.34.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
264
+ "model.layers.34.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
265
+ "model.layers.34.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
266
+ "model.layers.34.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
267
+ "model.layers.34.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
268
+ "model.layers.34.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
269
+ "model.layers.35.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
270
+ "model.layers.35.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
271
+ "model.layers.35.mlp.gate_proj.weight": "pytorch_model-00005-of-00007.bin",
272
+ "model.layers.35.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
273
+ "model.layers.35.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
274
+ "model.layers.35.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
275
+ "model.layers.35.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
276
+ "model.layers.35.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
277
+ "model.layers.35.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
278
+ "model.layers.36.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
279
+ "model.layers.36.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
280
+ "model.layers.36.mlp.gate_proj.weight": "pytorch_model-00005-of-00007.bin",
281
+ "model.layers.36.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
282
+ "model.layers.36.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
283
+ "model.layers.36.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
284
+ "model.layers.36.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
285
+ "model.layers.36.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
286
+ "model.layers.36.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
287
+ "model.layers.37.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
288
+ "model.layers.37.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
289
+ "model.layers.37.mlp.gate_proj.weight": "pytorch_model-00005-of-00007.bin",
290
+ "model.layers.37.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
291
+ "model.layers.37.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
292
+ "model.layers.37.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
293
+ "model.layers.37.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
294
+ "model.layers.37.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
295
+ "model.layers.37.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
296
+ "model.layers.38.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
297
+ "model.layers.38.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
298
+ "model.layers.38.mlp.gate_proj.weight": "pytorch_model-00005-of-00007.bin",
299
+ "model.layers.38.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
300
+ "model.layers.38.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
301
+ "model.layers.38.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
302
+ "model.layers.38.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
303
+ "model.layers.38.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
304
+ "model.layers.38.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
305
+ "model.layers.39.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
306
+ "model.layers.39.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
307
+ "model.layers.39.mlp.gate_proj.weight": "pytorch_model-00005-of-00007.bin",
308
+ "model.layers.39.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
309
+ "model.layers.39.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
310
+ "model.layers.39.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
311
+ "model.layers.39.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
312
+ "model.layers.39.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
313
+ "model.layers.39.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
314
+ "model.layers.4.input_layernorm.weight": "pytorch_model-00001-of-00007.bin",
315
+ "model.layers.4.mlp.down_proj.weight": "pytorch_model-00001-of-00007.bin",
316
+ "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00001-of-00007.bin",
317
+ "model.layers.4.mlp.up_proj.weight": "pytorch_model-00001-of-00007.bin",
318
+ "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00007.bin",
319
+ "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
320
+ "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00001-of-00007.bin",
321
+ "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
322
+ "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
323
+ "model.layers.40.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
324
+ "model.layers.40.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
325
+ "model.layers.40.mlp.gate_proj.weight": "pytorch_model-00005-of-00007.bin",
326
+ "model.layers.40.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
327
+ "model.layers.40.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
328
+ "model.layers.40.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
329
+ "model.layers.40.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
330
+ "model.layers.40.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
331
+ "model.layers.40.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
332
+ "model.layers.41.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
333
+ "model.layers.41.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
334
+ "model.layers.41.mlp.gate_proj.weight": "pytorch_model-00005-of-00007.bin",
335
+ "model.layers.41.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
336
+ "model.layers.41.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
337
+ "model.layers.41.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
338
+ "model.layers.41.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
339
+ "model.layers.41.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
340
+ "model.layers.41.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
341
+ "model.layers.42.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
342
+ "model.layers.42.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
343
+ "model.layers.42.mlp.gate_proj.weight": "pytorch_model-00005-of-00007.bin",
344
+ "model.layers.42.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
345
+ "model.layers.42.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
346
+ "model.layers.42.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
347
+ "model.layers.42.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
348
+ "model.layers.42.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
349
+ "model.layers.42.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
350
+ "model.layers.43.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
351
+ "model.layers.43.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
352
+ "model.layers.43.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
353
+ "model.layers.43.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
354
+ "model.layers.43.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
355
+ "model.layers.43.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
356
+ "model.layers.43.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
357
+ "model.layers.43.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
358
+ "model.layers.43.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
359
+ "model.layers.44.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
360
+ "model.layers.44.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
361
+ "model.layers.44.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
362
+ "model.layers.44.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
363
+ "model.layers.44.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
364
+ "model.layers.44.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
365
+ "model.layers.44.self_attn.o_proj.weight": "pytorch_model-00006-of-00007.bin",
366
+ "model.layers.44.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
367
+ "model.layers.44.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
368
+ "model.layers.45.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
369
+ "model.layers.45.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
370
+ "model.layers.45.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
371
+ "model.layers.45.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
372
+ "model.layers.45.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
373
+ "model.layers.45.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
374
+ "model.layers.45.self_attn.o_proj.weight": "pytorch_model-00006-of-00007.bin",
375
+ "model.layers.45.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
376
+ "model.layers.45.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
377
+ "model.layers.46.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
378
+ "model.layers.46.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
379
+ "model.layers.46.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
380
+ "model.layers.46.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
381
+ "model.layers.46.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
382
+ "model.layers.46.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
383
+ "model.layers.46.self_attn.o_proj.weight": "pytorch_model-00006-of-00007.bin",
384
+ "model.layers.46.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
385
+ "model.layers.46.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
386
+ "model.layers.47.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
387
+ "model.layers.47.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
388
+ "model.layers.47.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
389
+ "model.layers.47.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
390
+ "model.layers.47.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
391
+ "model.layers.47.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
392
+ "model.layers.47.self_attn.o_proj.weight": "pytorch_model-00006-of-00007.bin",
393
+ "model.layers.47.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
394
+ "model.layers.47.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
395
+ "model.layers.48.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
396
+ "model.layers.48.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
397
+ "model.layers.48.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
398
+ "model.layers.48.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
399
+ "model.layers.48.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
400
+ "model.layers.48.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
401
+ "model.layers.48.self_attn.o_proj.weight": "pytorch_model-00006-of-00007.bin",
402
+ "model.layers.48.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
403
+ "model.layers.48.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
404
+ "model.layers.49.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
405
+ "model.layers.49.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
406
+ "model.layers.49.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
407
+ "model.layers.49.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
408
+ "model.layers.49.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
409
+ "model.layers.49.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
410
+ "model.layers.49.self_attn.o_proj.weight": "pytorch_model-00006-of-00007.bin",
411
+ "model.layers.49.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
412
+ "model.layers.49.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
413
+ "model.layers.5.input_layernorm.weight": "pytorch_model-00001-of-00007.bin",
414
+ "model.layers.5.mlp.down_proj.weight": "pytorch_model-00001-of-00007.bin",
415
+ "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00001-of-00007.bin",
416
+ "model.layers.5.mlp.up_proj.weight": "pytorch_model-00001-of-00007.bin",
417
+ "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00007.bin",
418
+ "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
419
+ "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00001-of-00007.bin",
420
+ "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
421
+ "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
422
+ "model.layers.50.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
423
+ "model.layers.50.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
424
+ "model.layers.50.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
425
+ "model.layers.50.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
426
+ "model.layers.50.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
427
+ "model.layers.50.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
428
+ "model.layers.50.self_attn.o_proj.weight": "pytorch_model-00006-of-00007.bin",
429
+ "model.layers.50.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
430
+ "model.layers.50.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
431
+ "model.layers.51.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
432
+ "model.layers.51.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
433
+ "model.layers.51.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
434
+ "model.layers.51.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
435
+ "model.layers.51.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
436
+ "model.layers.51.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
437
+ "model.layers.51.self_attn.o_proj.weight": "pytorch_model-00006-of-00007.bin",
438
+ "model.layers.51.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
439
+ "model.layers.51.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
440
+ "model.layers.52.input_layernorm.weight": "pytorch_model-00007-of-00007.bin",
441
+ "model.layers.52.mlp.down_proj.weight": "pytorch_model-00007-of-00007.bin",
442
+ "model.layers.52.mlp.gate_proj.weight": "pytorch_model-00007-of-00007.bin",
443
+ "model.layers.52.mlp.up_proj.weight": "pytorch_model-00007-of-00007.bin",
444
+ "model.layers.52.post_attention_layernorm.weight": "pytorch_model-00007-of-00007.bin",
445
+ "model.layers.52.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
446
+ "model.layers.52.self_attn.o_proj.weight": "pytorch_model-00007-of-00007.bin",
447
+ "model.layers.52.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
448
+ "model.layers.52.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
449
+ "model.layers.53.input_layernorm.weight": "pytorch_model-00007-of-00007.bin",
450
+ "model.layers.53.mlp.down_proj.weight": "pytorch_model-00007-of-00007.bin",
451
+ "model.layers.53.mlp.gate_proj.weight": "pytorch_model-00007-of-00007.bin",
452
+ "model.layers.53.mlp.up_proj.weight": "pytorch_model-00007-of-00007.bin",
453
+ "model.layers.53.post_attention_layernorm.weight": "pytorch_model-00007-of-00007.bin",
454
+ "model.layers.53.self_attn.k_proj.weight": "pytorch_model-00007-of-00007.bin",
455
+ "model.layers.53.self_attn.o_proj.weight": "pytorch_model-00007-of-00007.bin",
456
+ "model.layers.53.self_attn.q_proj.weight": "pytorch_model-00007-of-00007.bin",
457
+ "model.layers.53.self_attn.v_proj.weight": "pytorch_model-00007-of-00007.bin",
458
+ "model.layers.54.input_layernorm.weight": "pytorch_model-00007-of-00007.bin",
459
+ "model.layers.54.mlp.down_proj.weight": "pytorch_model-00007-of-00007.bin",
460
+ "model.layers.54.mlp.gate_proj.weight": "pytorch_model-00007-of-00007.bin",
461
+ "model.layers.54.mlp.up_proj.weight": "pytorch_model-00007-of-00007.bin",
462
+ "model.layers.54.post_attention_layernorm.weight": "pytorch_model-00007-of-00007.bin",
463
+ "model.layers.54.self_attn.k_proj.weight": "pytorch_model-00007-of-00007.bin",
464
+ "model.layers.54.self_attn.o_proj.weight": "pytorch_model-00007-of-00007.bin",
465
+ "model.layers.54.self_attn.q_proj.weight": "pytorch_model-00007-of-00007.bin",
466
+ "model.layers.54.self_attn.v_proj.weight": "pytorch_model-00007-of-00007.bin",
467
+ "model.layers.55.input_layernorm.weight": "pytorch_model-00007-of-00007.bin",
468
+ "model.layers.55.mlp.down_proj.weight": "pytorch_model-00007-of-00007.bin",
469
+ "model.layers.55.mlp.gate_proj.weight": "pytorch_model-00007-of-00007.bin",
470
+ "model.layers.55.mlp.up_proj.weight": "pytorch_model-00007-of-00007.bin",
471
+ "model.layers.55.post_attention_layernorm.weight": "pytorch_model-00007-of-00007.bin",
472
+ "model.layers.55.self_attn.k_proj.weight": "pytorch_model-00007-of-00007.bin",
473
+ "model.layers.55.self_attn.o_proj.weight": "pytorch_model-00007-of-00007.bin",
474
+ "model.layers.55.self_attn.q_proj.weight": "pytorch_model-00007-of-00007.bin",
475
+ "model.layers.55.self_attn.v_proj.weight": "pytorch_model-00007-of-00007.bin",
476
+ "model.layers.56.input_layernorm.weight": "pytorch_model-00007-of-00007.bin",
477
+ "model.layers.56.mlp.down_proj.weight": "pytorch_model-00007-of-00007.bin",
478
+ "model.layers.56.mlp.gate_proj.weight": "pytorch_model-00007-of-00007.bin",
479
+ "model.layers.56.mlp.up_proj.weight": "pytorch_model-00007-of-00007.bin",
480
+ "model.layers.56.post_attention_layernorm.weight": "pytorch_model-00007-of-00007.bin",
481
+ "model.layers.56.self_attn.k_proj.weight": "pytorch_model-00007-of-00007.bin",
482
+ "model.layers.56.self_attn.o_proj.weight": "pytorch_model-00007-of-00007.bin",
483
+ "model.layers.56.self_attn.q_proj.weight": "pytorch_model-00007-of-00007.bin",
484
+ "model.layers.56.self_attn.v_proj.weight": "pytorch_model-00007-of-00007.bin",
485
+ "model.layers.57.input_layernorm.weight": "pytorch_model-00007-of-00007.bin",
486
+ "model.layers.57.mlp.down_proj.weight": "pytorch_model-00007-of-00007.bin",
487
+ "model.layers.57.mlp.gate_proj.weight": "pytorch_model-00007-of-00007.bin",
488
+ "model.layers.57.mlp.up_proj.weight": "pytorch_model-00007-of-00007.bin",
489
+ "model.layers.57.post_attention_layernorm.weight": "pytorch_model-00007-of-00007.bin",
490
+ "model.layers.57.self_attn.k_proj.weight": "pytorch_model-00007-of-00007.bin",
491
+ "model.layers.57.self_attn.o_proj.weight": "pytorch_model-00007-of-00007.bin",
492
+ "model.layers.57.self_attn.q_proj.weight": "pytorch_model-00007-of-00007.bin",
493
+ "model.layers.57.self_attn.v_proj.weight": "pytorch_model-00007-of-00007.bin",
494
+ "model.layers.58.input_layernorm.weight": "pytorch_model-00007-of-00007.bin",
495
+ "model.layers.58.mlp.down_proj.weight": "pytorch_model-00007-of-00007.bin",
496
+ "model.layers.58.mlp.gate_proj.weight": "pytorch_model-00007-of-00007.bin",
497
+ "model.layers.58.mlp.up_proj.weight": "pytorch_model-00007-of-00007.bin",
498
+ "model.layers.58.post_attention_layernorm.weight": "pytorch_model-00007-of-00007.bin",
499
+ "model.layers.58.self_attn.k_proj.weight": "pytorch_model-00007-of-00007.bin",
500
+ "model.layers.58.self_attn.o_proj.weight": "pytorch_model-00007-of-00007.bin",
501
+ "model.layers.58.self_attn.q_proj.weight": "pytorch_model-00007-of-00007.bin",
502
+ "model.layers.58.self_attn.v_proj.weight": "pytorch_model-00007-of-00007.bin",
503
+ "model.layers.59.input_layernorm.weight": "pytorch_model-00007-of-00007.bin",
504
+ "model.layers.59.mlp.down_proj.weight": "pytorch_model-00007-of-00007.bin",
505
+ "model.layers.59.mlp.gate_proj.weight": "pytorch_model-00007-of-00007.bin",
506
+ "model.layers.59.mlp.up_proj.weight": "pytorch_model-00007-of-00007.bin",
507
+ "model.layers.59.post_attention_layernorm.weight": "pytorch_model-00007-of-00007.bin",
508
+ "model.layers.59.self_attn.k_proj.weight": "pytorch_model-00007-of-00007.bin",
509
+ "model.layers.59.self_attn.o_proj.weight": "pytorch_model-00007-of-00007.bin",
510
+ "model.layers.59.self_attn.q_proj.weight": "pytorch_model-00007-of-00007.bin",
511
+ "model.layers.59.self_attn.v_proj.weight": "pytorch_model-00007-of-00007.bin",
512
+ "model.layers.6.input_layernorm.weight": "pytorch_model-00001-of-00007.bin",
513
+ "model.layers.6.mlp.down_proj.weight": "pytorch_model-00001-of-00007.bin",
514
+ "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00001-of-00007.bin",
515
+ "model.layers.6.mlp.up_proj.weight": "pytorch_model-00001-of-00007.bin",
516
+ "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00007.bin",
517
+ "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
518
+ "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00001-of-00007.bin",
519
+ "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
520
+ "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
521
+ "model.layers.7.input_layernorm.weight": "pytorch_model-00001-of-00007.bin",
522
+ "model.layers.7.mlp.down_proj.weight": "pytorch_model-00001-of-00007.bin",
523
+ "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00001-of-00007.bin",
524
+ "model.layers.7.mlp.up_proj.weight": "pytorch_model-00001-of-00007.bin",
525
+ "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00007.bin",
526
+ "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
527
+ "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00001-of-00007.bin",
528
+ "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
529
+ "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
530
+ "model.layers.8.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
531
+ "model.layers.8.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
532
+ "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
533
+ "model.layers.8.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
534
+ "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
535
+ "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
536
+ "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
537
+ "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
538
+ "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
539
+ "model.layers.9.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
540
+ "model.layers.9.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
541
+ "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
542
+ "model.layers.9.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
543
+ "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
544
+ "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00002-of-00007.bin",
545
+ "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
546
+ "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00002-of-00007.bin",
547
+ "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00002-of-00007.bin",
548
+ "model.norm.weight": "pytorch_model-00007-of-00007.bin"
549
+ }
550
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<|startoftext|>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<|endoftext|>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "<unk>",
18
+ "lstrip": false,
19
+ "normalized": true,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "unk_token": {
24
+ "content": "<unk>",
25
+ "lstrip": false,
26
+ "normalized": true,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ }
30
+ }
tokenization_yi.py ADDED
@@ -0,0 +1,255 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ from shutil import copyfile
3
+ from typing import Any, Dict, List, Optional, Tuple
4
+
5
+ import sentencepiece as spm
6
+ from transformers.tokenization_utils import AddedToken, PreTrainedTokenizer
7
+ from transformers.utils import logging
8
+
9
+ logger = logging.get_logger(__name__)
10
+
11
+ VOCAB_FILES_NAMES = {"vocab_file": "tokenizer.model"}
12
+
13
+ PRETRAINED_VOCAB_FILES_MAP = {
14
+ "vocab_file": {},
15
+ "tokenizer_file": {},
16
+ }
17
+ PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {}
18
+
19
+
20
+ class YiTokenizer(PreTrainedTokenizer):
21
+ """
22
+ Construct a Yi tokenizer. Based on byte-level Byte-Pair-Encoding.
23
+
24
+ Args:
25
+ vocab_file (`str`):
26
+ Path to the vocabulary file.
27
+ """
28
+
29
+ vocab_files_names = VOCAB_FILES_NAMES
30
+ pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
31
+ max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
32
+ model_input_names = ["input_ids", "attention_mask"]
33
+
34
+ def __init__(
35
+ self,
36
+ vocab_file,
37
+ unk_token="<unk>",
38
+ bos_token="<|startoftext|>",
39
+ eos_token="<|endoftext|>",
40
+ pad_token="<unk>",
41
+ sp_model_kwargs: Optional[Dict[str, Any]] = None,
42
+ add_bos_token=True,
43
+ add_eos_token=False,
44
+ clean_up_tokenization_spaces=False,
45
+ **kwargs,
46
+ ):
47
+ self.sp_model_kwargs = {} if sp_model_kwargs is None else sp_model_kwargs
48
+ bos_token = (
49
+ AddedToken(bos_token, lstrip=False, rstrip=False)
50
+ if isinstance(bos_token, str)
51
+ else bos_token
52
+ )
53
+ eos_token = (
54
+ AddedToken(eos_token, lstrip=False, rstrip=False)
55
+ if isinstance(eos_token, str)
56
+ else eos_token
57
+ )
58
+ unk_token = (
59
+ AddedToken(unk_token, lstrip=False, rstrip=False)
60
+ if isinstance(unk_token, str)
61
+ else unk_token
62
+ )
63
+ pad_token = (
64
+ AddedToken(pad_token, lstrip=False, rstrip=False)
65
+ if isinstance(pad_token, str)
66
+ else pad_token
67
+ )
68
+ self.vocab_file = vocab_file
69
+ self.add_bos_token = add_bos_token
70
+ self.add_eos_token = add_eos_token
71
+ self.sp_model = spm.SentencePieceProcessor(**self.sp_model_kwargs)
72
+ self.sp_model.Load(vocab_file)
73
+ super().__init__(
74
+ bos_token=bos_token,
75
+ eos_token=eos_token,
76
+ unk_token=unk_token,
77
+ pad_token=pad_token,
78
+ add_bos_token=add_bos_token,
79
+ add_eos_token=add_eos_token,
80
+ sp_model_kwargs=self.sp_model_kwargs,
81
+ clean_up_tokenization_spaces=clean_up_tokenization_spaces,
82
+ **kwargs,
83
+ )
84
+
85
+ def __getstate__(self):
86
+ state = self.__dict__.copy()
87
+ state["sp_model"] = None
88
+ return state
89
+
90
+ def __setstate__(self, d):
91
+ self.__dict__ = d
92
+ self.sp_model = spm.SentencePieceProcessor(**self.sp_model_kwargs)
93
+ self.sp_model.Load(self.vocab_file)
94
+
95
+ @property
96
+ def vocab_size(self):
97
+ """Returns vocab size"""
98
+ return self.sp_model.get_piece_size()
99
+
100
+ def get_vocab(self):
101
+ """Returns vocab as a dict"""
102
+ vocab = {self.convert_ids_to_tokens(i): i for i in range(self.vocab_size)}
103
+ vocab.update(self.added_tokens_encoder)
104
+ return vocab
105
+
106
+ def _tokenize(self, text):
107
+ """Returns a tokenized string."""
108
+ return self.sp_model.encode(text, out_type=str)
109
+
110
+ def _convert_token_to_id(self, token):
111
+ """Converts a token (str) in an id using the vocab."""
112
+ return self.sp_model.piece_to_id(token)
113
+
114
+ def _convert_id_to_token(self, index):
115
+ """Converts an index (integer) in a token (str) using the vocab."""
116
+ token = self.sp_model.IdToPiece(index)
117
+ return token
118
+
119
+ def convert_tokens_to_string(self, tokens):
120
+ """Converts a sequence of tokens (string) in a single string."""
121
+ current_sub_tokens = []
122
+ out_string = ""
123
+ prev_is_special = False
124
+ for i, token in enumerate(tokens):
125
+ # make sure that special tokens are not decoded using sentencepiece model
126
+ if token in self.all_special_tokens:
127
+ if not prev_is_special and i != 0:
128
+ out_string += " "
129
+ out_string += self.sp_model.decode(current_sub_tokens) + token
130
+ prev_is_special = True
131
+ current_sub_tokens = []
132
+ else:
133
+ current_sub_tokens.append(token)
134
+ prev_is_special = False
135
+ out_string += self.sp_model.decode(current_sub_tokens)
136
+ return out_string
137
+
138
+ def save_vocabulary(
139
+ self, save_directory, filename_prefix: Optional[str] = None
140
+ ) -> Tuple[str]:
141
+ """
142
+ Save the vocabulary and special tokens file to a directory.
143
+
144
+ Args:
145
+ save_directory (`str`):
146
+ The directory in which to save the vocabulary.
147
+
148
+ Returns:
149
+ `Tuple(str)`: Paths to the files saved.
150
+ """
151
+ if not os.path.isdir(save_directory):
152
+ logger.error(f"Vocabulary path ({save_directory}) should be a directory")
153
+ return
154
+ out_vocab_file = os.path.join(
155
+ save_directory,
156
+ (filename_prefix + "-" if filename_prefix else "")
157
+ + VOCAB_FILES_NAMES["vocab_file"],
158
+ )
159
+
160
+ if os.path.abspath(self.vocab_file) != os.path.abspath(
161
+ out_vocab_file
162
+ ) and os.path.isfile(self.vocab_file):
163
+ copyfile(self.vocab_file, out_vocab_file)
164
+ elif not os.path.isfile(self.vocab_file):
165
+ with open(out_vocab_file, "wb") as fi:
166
+ content_spiece_model = self.sp_model.serialized_model_proto()
167
+ fi.write(content_spiece_model)
168
+
169
+ return (out_vocab_file,)
170
+
171
+ def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
172
+ bos_token_id = [self.bos_token_id] if self.add_bos_token else []
173
+ eos_token_id = [self.eos_token_id] if self.add_eos_token else []
174
+
175
+ output = bos_token_id + token_ids_0 + eos_token_id
176
+
177
+ if token_ids_1 is not None:
178
+ output = output + bos_token_id + token_ids_1 + eos_token_id
179
+
180
+ return output
181
+
182
+ def get_special_tokens_mask(
183
+ self,
184
+ token_ids_0: List[int],
185
+ token_ids_1: Optional[List[int]] = None,
186
+ already_has_special_tokens: bool = False,
187
+ ) -> List[int]:
188
+ """
189
+ Retrieve sequence ids from a token list that has no special tokens added. This method is called when adding
190
+ special tokens using the tokenizer `prepare_for_model` method.
191
+
192
+ Args:
193
+ token_ids_0 (`List[int]`):
194
+ List of IDs.
195
+ token_ids_1 (`List[int]`, *optional*):
196
+ Optional second list of IDs for sequence pairs.
197
+ already_has_special_tokens (`bool`, *optional*, defaults to `False`):
198
+ Whether or not the token list is already formatted with special tokens for the model.
199
+
200
+ Returns:
201
+ `List[int]`: A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token.
202
+ """
203
+ if already_has_special_tokens:
204
+ return super().get_special_tokens_mask(
205
+ token_ids_0=token_ids_0,
206
+ token_ids_1=token_ids_1,
207
+ already_has_special_tokens=True,
208
+ )
209
+
210
+ bos_token_id = [1] if self.add_bos_token else []
211
+ eos_token_id = [1] if self.add_eos_token else []
212
+
213
+ if token_ids_1 is None:
214
+ return bos_token_id + ([0] * len(token_ids_0)) + eos_token_id
215
+ return (
216
+ bos_token_id
217
+ + ([0] * len(token_ids_0))
218
+ + eos_token_id
219
+ + bos_token_id
220
+ + ([0] * len(token_ids_1))
221
+ + eos_token_id
222
+ )
223
+
224
+ def create_token_type_ids_from_sequences(
225
+ self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
226
+ ) -> List[int]:
227
+ """
228
+ Creates a mask from the two sequences passed to be used in a sequence-pair classification task. An ALBERT
229
+ sequence pair mask has the following format:
230
+
231
+ ```
232
+ 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1
233
+ | first sequence | second sequence |
234
+ ```
235
+
236
+ if token_ids_1 is None, only returns the first portion of the mask (0s).
237
+
238
+ Args:
239
+ token_ids_0 (`List[int]`):
240
+ List of ids.
241
+ token_ids_1 (`List[int]`, *optional*):
242
+ Optional second list of IDs for sequence pairs.
243
+
244
+ Returns:
245
+ `List[int]`: List of [token type IDs](../glossary#token-type-ids) according to the given sequence(s).
246
+ """
247
+ bos_token_id = [self.bos_token_id] if self.add_bos_token else []
248
+ eos_token_id = [self.eos_token_id] if self.add_eos_token else []
249
+
250
+ output = [0] * len(bos_token_id + token_ids_0 + eos_token_id)
251
+
252
+ if token_ids_1 is not None:
253
+ output += [1] * len(bos_token_id + token_ids_1 + eos_token_id)
254
+
255
+ return output
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:386c49cf943d71aa110361135338c50e38beeff0a66593480421f37b319e1a39
3
+ size 1033105
tokenizer_config.json ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_eos_token": false,
4
+ "added_tokens_decoder": {
5
+ "0": {
6
+ "content": "<unk>",
7
+ "lstrip": false,
8
+ "normalized": true,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "1": {
14
+ "content": "<|startoftext|>",
15
+ "lstrip": false,
16
+ "normalized": true,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "2": {
22
+ "content": "<|endoftext|>",
23
+ "lstrip": false,
24
+ "normalized": true,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ }
29
+ },
30
+ "auto_map": {
31
+ "AutoTokenizer": [
32
+ "tokenization_yi.YiTokenizer",
33
+ null
34
+ ]
35
+ },
36
+ "bos_token": "<|startoftext|>",
37
+ "clean_up_tokenization_spaces": false,
38
+ "eos_token": "<|endoftext|>",
39
+ "model_max_length": 200000,
40
+ "pad_token": "<unk>",
41
+ "sp_model_kwargs": {},
42
+ "tokenizer_class": "YiTokenizer",
43
+ "unk_token": "<unk>"
44
+ }