Add pipeline tag and link to project page
Browse filesThis PR improves the model card by:
- Adding the relevant `pipeline_tag`, ensuring people can find your model at https://huggingface.co/models?pipeline_tag=text-generation
- Adding a link to the project page at https://abdelfattah-lab.github.io/TokenButler/
README.md
CHANGED
@@ -1,9 +1,11 @@
|
|
1 |
---
|
2 |
-
license: mit
|
3 |
-
library_name: transformers
|
4 |
base_model:
|
5 |
- deepseek-ai/DeepSeek-R1-Distill-Llama-8B
|
|
|
|
|
|
|
6 |
---
|
|
|
7 |
# TokenButler
|
8 |
<!-- markdownlint-disable first-line-h1 -->
|
9 |
<!-- markdownlint-disable html -->
|
@@ -17,7 +19,7 @@ base_model:
|
|
17 |
<hr>
|
18 |
<div align="center" style="line-height: 1;">
|
19 |
<!-- Paper Badge -->
|
20 |
-
<a href="https://
|
21 |
<img alt="Paper"
|
22 |
src="https://img.shields.io/badge/Paper-View-orange?logo=readthedocs&logoColor=white"
|
23 |
style="display: inline-block; vertical-align: middle;"/>
|
@@ -28,6 +30,12 @@ base_model:
|
|
28 |
src="https://img.shields.io/badge/GitHub-Repo-black?logo=github&logoColor=white"
|
29 |
style="display: inline-block; vertical-align: middle;"/>
|
30 |
</a>
|
|
|
|
|
|
|
|
|
|
|
|
|
31 |
</div>
|
32 |
|
33 |
<br>
|
@@ -72,4 +80,121 @@ model = set_sparsity(model, "fixed_60pc")
|
|
72 |
# Custom Synthetic Task
|
73 |
<div align="center">
|
74 |
<img src="https://github.com/abdelfattah-lab/TokenButler/blob/main/figs/datasetfig.png?raw=true" width="100%" alt="Synthetic Tasks" />
|
75 |
-
</div>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
|
|
|
|
2 |
base_model:
|
3 |
- deepseek-ai/DeepSeek-R1-Distill-Llama-8B
|
4 |
+
library_name: transformers
|
5 |
+
license: mit
|
6 |
+
pipeline_tag: text-generation
|
7 |
---
|
8 |
+
|
9 |
# TokenButler
|
10 |
<!-- markdownlint-disable first-line-h1 -->
|
11 |
<!-- markdownlint-disable html -->
|
|
|
19 |
<hr>
|
20 |
<div align="center" style="line-height: 1;">
|
21 |
<!-- Paper Badge -->
|
22 |
+
<a href="https://arxiv.org/abs/2503.07518" target="_blank" style="margin: 2px;">
|
23 |
<img alt="Paper"
|
24 |
src="https://img.shields.io/badge/Paper-View-orange?logo=readthedocs&logoColor=white"
|
25 |
style="display: inline-block; vertical-align: middle;"/>
|
|
|
30 |
src="https://img.shields.io/badge/GitHub-Repo-black?logo=github&logoColor=white"
|
31 |
style="display: inline-block; vertical-align: middle;"/>
|
32 |
</a>
|
33 |
+
<!-- Project Page Badge -->
|
34 |
+
<a href="https://abdelfattah-lab.github.io/TokenButler/" target="_blank" style="margin: 2px;">
|
35 |
+
<img alt="Project Page"
|
36 |
+
src="https://img.shields.io/badge/Project%20Page-🌐-lightgrey"
|
37 |
+
style="display: inline-block; vertical-align: middle;"/>
|
38 |
+
</a>
|
39 |
</div>
|
40 |
|
41 |
<br>
|
|
|
80 |
# Custom Synthetic Task
|
81 |
<div align="center">
|
82 |
<img src="https://github.com/abdelfattah-lab/TokenButler/blob/main/figs/datasetfig.png?raw=true" width="100%" alt="Synthetic Tasks" />
|
83 |
+
</div>
|
84 |
+
|
85 |
+
# File information
|
86 |
+
|
87 |
+
The repository contains the following file information:
|
88 |
+
|
89 |
+
Filename: tokenizer.json
|
90 |
+
Content: "Content of the file is larger than 50 KB, too long to display."
|
91 |
+
|
92 |
+
Filename: pytorch_model.bin.index.json
|
93 |
+
Content: "Content of the file is larger than 50 KB, too long to display."
|
94 |
+
|
95 |
+
Filename: generation_config.json
|
96 |
+
Content: {
|
97 |
+
"_from_model_config": true,
|
98 |
+
"bos_token_id": 128000,
|
99 |
+
"eos_token_id": 128001,
|
100 |
+
"transformers_version": "4.48.3"
|
101 |
+
}
|
102 |
+
|
103 |
+
Filename: tokenizer_config.json
|
104 |
+
Content: {
|
105 |
+
"add_bos_token": true,
|
106 |
+
"add_eos_token": false,
|
107 |
+
"bos_token": {
|
108 |
+
"__type": "AddedToken",
|
109 |
+
"content": "<\uff5cbegin\u2581of\u2581sentence\uff5c>",
|
110 |
+
"lstrip": false,
|
111 |
+
"normalized": true,
|
112 |
+
"rstrip": false,
|
113 |
+
"single_word": false
|
114 |
+
},
|
115 |
+
"clean_up_tokenization_spaces": false,
|
116 |
+
"eos_token": {
|
117 |
+
"__type": "AddedToken",
|
118 |
+
"content": "<\uff5cend\u2581of\u2581sentence\uff5c>",
|
119 |
+
"lstrip": false,
|
120 |
+
"normalized": true,
|
121 |
+
"rstrip": false,
|
122 |
+
"single_word": false
|
123 |
+
},
|
124 |
+
"legacy": true,
|
125 |
+
"model_max_length": 16384,
|
126 |
+
"pad_token": {
|
127 |
+
"__type": "AddedToken",
|
128 |
+
"content": "<\uff5cend\u2581of\u2581sentence\uff5c>",
|
129 |
+
"lstrip": false,
|
130 |
+
"normalized": true,
|
131 |
+
"rstrip": false,
|
132 |
+
"single_word": false
|
133 |
+
},
|
134 |
+
"sp_model_kwargs": {},
|
135 |
+
"unk_token": null,
|
136 |
+
"tokenizer_class": "LlamaTokenizerFast",
|
137 |
+
"chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set ns = namespace(is_first=false, is_tool=false, is_output_first=true, system_prompt='') %}{%- for message in messages %}{%- if message['role'] == 'system' %}{% set ns.system_prompt = message['content'] %}{%- endif %}{%- endfor %}{{bos_token}}{{ns.system_prompt}}{%- for message in messages %}{%- if message['role'] == 'user' %}{%- set ns.is_tool = false -%}{{'<\uff5cUser\uff5c>' + message['content']}}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is none %}{%- set ns.is_tool = false -%}{%- for tool in message['tool_calls']%}{%- if not ns.is_first %}{{'<\uff5cAssistant\uff5c><\uff5ctool\u2581calls\u2581begin\uff5c><\uff5ctool\u2581call\u2581begin\uff5c>' + tool['type'] + '<\uff5ctool\u2581sep\uff5c>' + tool['function']['name'] + '\
|
138 |
+
' + '```json' + '\
|
139 |
+
' + tool['function']['arguments'] + '\
|
140 |
+
' + '```' + '<\uff5ctool\u2581call\u2581end\uff5c>'}}{%- set ns.is_first = true -%}{%- else %}{{'\
|
141 |
+
' + '<\uff5ctool\u2581call\u2581begin\uff5c>' + tool['type'] + '<\uff5ctool\u2581sep\uff5c>' + tool['function']['name'] + '\
|
142 |
+
' + '```json' + '\
|
143 |
+
' + tool['function']['arguments'] + '\
|
144 |
+
' + '```' + '<\uff5ctool\u2581call\u2581end\uff5c>'}}{{'<\uff5ctool\u2581calls\u2581end\uff5c><\uff5cend\u2581of\u2581sentence\uff5c>'}}{%- endif %}{%- endfor %}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is not none %}{%- if ns.is_tool %}{{'<\uff5ctool\u2581outputs\u2581end\uff5c>' + message['content'] + '<\uff5cend\u2581of\u2581sentence\uff5c>'}}{%- set ns.is_tool = false -%}{%- else %}{% set content = message['content'] %}{% if '</think>' in content %}{% set content = content.split('</think>')[-1] %}{{'<\uff5cAssistant\uff5c>' + content + '<\uff5cend\u2581of\u2581sentence\uff5c>'}}{%- endif %}{%- endif %}{%- endif %}{%- if message['role'] == 'tool' %}{%- set ns.is_tool = true -%}{%- if ns.is_output_first %}{{'<\uff5ctool\u2581outputs\u2581begin\uff5c><\uff5ctool\u2581output\u2581begin\uff5c>' + message['content'] + '<\uff5ctool\u2581output\u2581end\uff5c>'}}{%- set ns.is_output_first = false %}{%- else %}{{'\
|
145 |
+
<\uff5ctool\u2581output\u2581begin\uff5c>' + message['content'] + '<\uff5ctool\u2581output\u2581end\uff5c>'}}{%- endif %}{%- endif %}{%- endfor -%}{% if ns.is_tool %}{{'<\uff5ctool\u2581outputs\u2581end\uff5c>'}}{% endif %}{% if add_generation_prompt and not ns.is_tool %}{{'<\uff5cAssistant\uff5c><think>\
|
146 |
+
'}}{% endif %}"
|
147 |
+
}
|
148 |
+
|
149 |
+
Filename: config.json
|
150 |
+
Content: {
|
151 |
+
"architectures": [
|
152 |
+
"modeling_llama_butler.LlamaButlerForCausalLM"
|
153 |
+
],
|
154 |
+
"attention_bias": false,
|
155 |
+
"attention_dropout": 0.0,
|
156 |
+
"attn_reduce_factor": 8,
|
157 |
+
"auto_map": {
|
158 |
+
"AutoConfig": "modeling_llama_butler.LlamaButlerConfig",
|
159 |
+
"AutoModel": "modeling_llama_butler.LlamaButlerForCausalLM",
|
160 |
+
"AutoModelForCausalLM": "modeling_llama_butler.LlamaButlerForCausalLM"
|
161 |
+
},
|
162 |
+
"bos_token_id": 128000,
|
163 |
+
"dDash": 32,
|
164 |
+
"eos_token_id": 128001,
|
165 |
+
"eval_llm_mode": "ExpPred",
|
166 |
+
"flash_attn": false,
|
167 |
+
"head_attn_reduce_factor": 2,
|
168 |
+
"head_dim": 128,
|
169 |
+
"hidden_act": "silu",
|
170 |
+
"hidden_size": 4096,
|
171 |
+
"initializer_range": 0.02,
|
172 |
+
"intdim": 1024,
|
173 |
+
"intermediate_size": 14336,
|
174 |
+
"lookahead": 0,
|
175 |
+
"max_position_embeddings": 131072,
|
176 |
+
"min_sparse_index": 8,
|
177 |
+
"mlp_bias": false,
|
178 |
+
"model_type": "llama_butler",
|
179 |
+
"num_attention_heads": 32,
|
180 |
+
"num_hidden_layers": 32,
|
181 |
+
"num_key_value_heads": 8,
|
182 |
+
"pretraining_tp": 1,
|
183 |
+
"producer_frequency": 32,
|
184 |
+
"rms_norm_eps": 1e-05,
|
185 |
+
"rope_scaling": {
|
186 |
+
"factor": 8.0,
|
187 |
+
"high_freq_factor": 4.0,
|
188 |
+
"low_freq_factor": 1.0,
|
189 |
+
"original_max_position_embeddings": 8192,
|
190 |
+
"rope_type": "llama3"
|
191 |
+
},
|
192 |
+
"rope_theta": 500000.0,
|
193 |
+
"sliding_window": 128,
|
194 |
+
"token_sparse_method": "fixed_50pc",
|
195 |
+
"torch_dtype": "float32",
|
196 |
+
"train_headpredictor": false,
|
197 |
+
"transformers_version": "4.48.3",
|
198 |
+
"use_cache": true,
|
199 |
+
"vocab_size": 128256
|
200 |
+
}
|