Commit
•
4b8f2b6
1
Parent(s):
19afde9
Upload folder using huggingface_hub
Browse files- README.md +198 -0
- config.json +28 -0
- generation_config.json +7 -0
- output.safetensors +3 -0
- special_tokens_map.json +30 -0
- tokenizer.model +3 -0
- tokenizer_config.json +44 -0
README.md
ADDED
@@ -0,0 +1,198 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- en
|
4 |
+
license: apache-2.0
|
5 |
+
tags:
|
6 |
+
- smol_llama
|
7 |
+
- llama2
|
8 |
+
datasets:
|
9 |
+
- JeanKaddour/minipile
|
10 |
+
- pszemraj/simple_wikipedia_LM
|
11 |
+
- mattymchen/refinedweb-3m
|
12 |
+
- BEE-spoke-data/knowledge-inoc-concat-v1
|
13 |
+
inference:
|
14 |
+
parameters:
|
15 |
+
max_new_tokens: 64
|
16 |
+
do_sample: true
|
17 |
+
temperature: 0.8
|
18 |
+
repetition_penalty: 1.05
|
19 |
+
no_repeat_ngram_size: 4
|
20 |
+
eta_cutoff: 0.0006
|
21 |
+
renormalize_logits: true
|
22 |
+
widget:
|
23 |
+
- text: My name is El Microondas the Wise, and
|
24 |
+
example_title: El Microondas
|
25 |
+
- text: Kennesaw State University is a public
|
26 |
+
example_title: Kennesaw State University
|
27 |
+
- text: Bungie Studios is an American video game developer. They are most famous for
|
28 |
+
developing the award winning Halo series of video games. They also made Destiny.
|
29 |
+
The studio was founded
|
30 |
+
example_title: Bungie
|
31 |
+
- text: The Mona Lisa is a world-renowned painting created by
|
32 |
+
example_title: Mona Lisa
|
33 |
+
- text: The Harry Potter series, written by J.K. Rowling, begins with the book titled
|
34 |
+
example_title: Harry Potter Series
|
35 |
+
- text: 'Question: I have cities, but no houses. I have mountains, but no trees. I
|
36 |
+
have water, but no fish. What am I?
|
37 |
+
|
38 |
+
Answer:'
|
39 |
+
example_title: Riddle
|
40 |
+
- text: The process of photosynthesis involves the conversion of
|
41 |
+
example_title: Photosynthesis
|
42 |
+
- text: Jane went to the store to buy some groceries. She picked up apples, oranges,
|
43 |
+
and a loaf of bread. When she got home, she realized she forgot
|
44 |
+
example_title: Story Continuation
|
45 |
+
- text: 'Problem 2: If a train leaves Station A at 9:00 AM and travels at 60 mph,
|
46 |
+
and another train leaves Station B at 10:00 AM and travels at 80 mph, when will
|
47 |
+
they meet if the distance between the stations is 300 miles?
|
48 |
+
|
49 |
+
To determine'
|
50 |
+
example_title: Math Problem
|
51 |
+
- text: In the context of computer programming, an algorithm is
|
52 |
+
example_title: Algorithm Definition
|
53 |
+
pipeline_tag: text-generation
|
54 |
+
model-index:
|
55 |
+
- name: smol_llama-220M-GQA
|
56 |
+
results:
|
57 |
+
- task:
|
58 |
+
type: text-generation
|
59 |
+
name: Text Generation
|
60 |
+
dataset:
|
61 |
+
name: AI2 Reasoning Challenge (25-Shot)
|
62 |
+
type: ai2_arc
|
63 |
+
config: ARC-Challenge
|
64 |
+
split: test
|
65 |
+
args:
|
66 |
+
num_few_shot: 25
|
67 |
+
metrics:
|
68 |
+
- type: acc_norm
|
69 |
+
value: 24.83
|
70 |
+
name: normalized accuracy
|
71 |
+
source:
|
72 |
+
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=BEE-spoke-data/smol_llama-220M-GQA
|
73 |
+
name: Open LLM Leaderboard
|
74 |
+
- task:
|
75 |
+
type: text-generation
|
76 |
+
name: Text Generation
|
77 |
+
dataset:
|
78 |
+
name: HellaSwag (10-Shot)
|
79 |
+
type: hellaswag
|
80 |
+
split: validation
|
81 |
+
args:
|
82 |
+
num_few_shot: 10
|
83 |
+
metrics:
|
84 |
+
- type: acc_norm
|
85 |
+
value: 29.76
|
86 |
+
name: normalized accuracy
|
87 |
+
source:
|
88 |
+
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=BEE-spoke-data/smol_llama-220M-GQA
|
89 |
+
name: Open LLM Leaderboard
|
90 |
+
- task:
|
91 |
+
type: text-generation
|
92 |
+
name: Text Generation
|
93 |
+
dataset:
|
94 |
+
name: MMLU (5-Shot)
|
95 |
+
type: cais/mmlu
|
96 |
+
config: all
|
97 |
+
split: test
|
98 |
+
args:
|
99 |
+
num_few_shot: 5
|
100 |
+
metrics:
|
101 |
+
- type: acc
|
102 |
+
value: 25.85
|
103 |
+
name: accuracy
|
104 |
+
source:
|
105 |
+
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=BEE-spoke-data/smol_llama-220M-GQA
|
106 |
+
name: Open LLM Leaderboard
|
107 |
+
- task:
|
108 |
+
type: text-generation
|
109 |
+
name: Text Generation
|
110 |
+
dataset:
|
111 |
+
name: TruthfulQA (0-shot)
|
112 |
+
type: truthful_qa
|
113 |
+
config: multiple_choice
|
114 |
+
split: validation
|
115 |
+
args:
|
116 |
+
num_few_shot: 0
|
117 |
+
metrics:
|
118 |
+
- type: mc2
|
119 |
+
value: 44.55
|
120 |
+
source:
|
121 |
+
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=BEE-spoke-data/smol_llama-220M-GQA
|
122 |
+
name: Open LLM Leaderboard
|
123 |
+
- task:
|
124 |
+
type: text-generation
|
125 |
+
name: Text Generation
|
126 |
+
dataset:
|
127 |
+
name: Winogrande (5-shot)
|
128 |
+
type: winogrande
|
129 |
+
config: winogrande_xl
|
130 |
+
split: validation
|
131 |
+
args:
|
132 |
+
num_few_shot: 5
|
133 |
+
metrics:
|
134 |
+
- type: acc
|
135 |
+
value: 50.99
|
136 |
+
name: accuracy
|
137 |
+
source:
|
138 |
+
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=BEE-spoke-data/smol_llama-220M-GQA
|
139 |
+
name: Open LLM Leaderboard
|
140 |
+
- task:
|
141 |
+
type: text-generation
|
142 |
+
name: Text Generation
|
143 |
+
dataset:
|
144 |
+
name: GSM8k (5-shot)
|
145 |
+
type: gsm8k
|
146 |
+
config: main
|
147 |
+
split: test
|
148 |
+
args:
|
149 |
+
num_few_shot: 5
|
150 |
+
metrics:
|
151 |
+
- type: acc
|
152 |
+
value: 0.68
|
153 |
+
name: accuracy
|
154 |
+
source:
|
155 |
+
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=BEE-spoke-data/smol_llama-220M-GQA
|
156 |
+
name: Open LLM Leaderboard
|
157 |
+
---
|
158 |
+
|
159 |
+
|
160 |
+
# smol_llama: 220M GQA
|
161 |
+
|
162 |
+
> model card WIP, more details to come
|
163 |
+
|
164 |
+
|
165 |
+
A small 220M param (total) decoder model. This is the first version of the model.
|
166 |
+
|
167 |
+
- 1024 hidden size, 10 layers
|
168 |
+
- GQA (32 heads, 8 key-value), context length 2048
|
169 |
+
- train-from-scratch on one GPU :)
|
170 |
+
|
171 |
+
## Links
|
172 |
+
|
173 |
+
[Here](https://huggingface.co/collections/BEE-spoke-data/finetuned-smol-220m-65998b080ae723e79c830f83) are some fine-tunes we did, but there are many more possibilities out there!
|
174 |
+
|
175 |
+
- instruct
|
176 |
+
- openhermes - [link](https://huggingface.co/BEE-spoke-data/smol_llama-220M-openhermes)
|
177 |
+
- open-instruct - [link](https://huggingface.co/BEE-spoke-data/smol_llama-220M-open_instruct)
|
178 |
+
- code
|
179 |
+
- python (pypi) - [link](https://huggingface.co/BEE-spoke-data/beecoder-220M-python)
|
180 |
+
- zephyr DPO tune
|
181 |
+
- SFT - [link](https://huggingface.co/BEE-spoke-data/zephyr-220m-sft-full)
|
182 |
+
- full DPO - [link](https://huggingface.co/BEE-spoke-data/zephyr-220m-dpo-full)
|
183 |
+
|
184 |
+
---
|
185 |
+
|
186 |
+
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
|
187 |
+
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_BEE-spoke-data__smol_llama-220M-GQA)
|
188 |
+
|
189 |
+
| Metric |Value|
|
190 |
+
|---------------------------------|----:|
|
191 |
+
|Avg. |29.44|
|
192 |
+
|AI2 Reasoning Challenge (25-Shot)|24.83|
|
193 |
+
|HellaSwag (10-Shot) |29.76|
|
194 |
+
|MMLU (5-Shot) |25.85|
|
195 |
+
|TruthfulQA (0-shot) |44.55|
|
196 |
+
|Winogrande (5-shot) |50.99|
|
197 |
+
|GSM8k (5-shot) | 0.68|
|
198 |
+
|
config.json
ADDED
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "BEE-spoke-data/NanoLlama-GQA-L10-A32_KV8-v17-KI",
|
3 |
+
"architectures": [
|
4 |
+
"LlamaForCausalLM"
|
5 |
+
],
|
6 |
+
"attention_bias": false,
|
7 |
+
"attention_dropout": 0.0,
|
8 |
+
"bos_token_id": 1,
|
9 |
+
"eos_token_id": 2,
|
10 |
+
"hidden_act": "silu",
|
11 |
+
"hidden_size": 1024,
|
12 |
+
"initializer_range": 0.02,
|
13 |
+
"intermediate_size": 4096,
|
14 |
+
"max_position_embeddings": 2048,
|
15 |
+
"model_type": "llama",
|
16 |
+
"num_attention_heads": 32,
|
17 |
+
"num_hidden_layers": 10,
|
18 |
+
"num_key_value_heads": 8,
|
19 |
+
"pretraining_tp": 1,
|
20 |
+
"rms_norm_eps": 1e-06,
|
21 |
+
"rope_scaling": null,
|
22 |
+
"rope_theta": 10000.0,
|
23 |
+
"tie_word_embeddings": false,
|
24 |
+
"torch_dtype": "bfloat16",
|
25 |
+
"transformers_version": "4.37.0.dev0",
|
26 |
+
"use_cache": false,
|
27 |
+
"vocab_size": 32128
|
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 |
+
"transformers_version": "4.37.0.dev0",
|
6 |
+
"use_cache": false
|
7 |
+
}
|
output.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b20fc7c1013c35474cc64a9fd93e30545b4e1383d35cbd66c59e1ec6cd78e307
|
3 |
+
size 180025464
|
special_tokens_map.json
ADDED
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "<s>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": false,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"eos_token": {
|
10 |
+
"content": "</s>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": false,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"pad_token": {
|
17 |
+
"content": "</s>",
|
18 |
+
"lstrip": false,
|
19 |
+
"normalized": false,
|
20 |
+
"rstrip": false,
|
21 |
+
"single_word": false
|
22 |
+
},
|
23 |
+
"unk_token": {
|
24 |
+
"content": "<unk>",
|
25 |
+
"lstrip": false,
|
26 |
+
"normalized": false,
|
27 |
+
"rstrip": false,
|
28 |
+
"single_word": false
|
29 |
+
}
|
30 |
+
}
|
tokenizer.model
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
|
3 |
+
size 499723
|
tokenizer_config.json
ADDED
@@ -0,0 +1,44 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_bos_token": true,
|
3 |
+
"add_eos_token": false,
|
4 |
+
"added_tokens_decoder": {
|
5 |
+
"0": {
|
6 |
+
"content": "<unk>",
|
7 |
+
"lstrip": false,
|
8 |
+
"normalized": false,
|
9 |
+
"rstrip": false,
|
10 |
+
"single_word": false,
|
11 |
+
"special": true
|
12 |
+
},
|
13 |
+
"1": {
|
14 |
+
"content": "<s>",
|
15 |
+
"lstrip": false,
|
16 |
+
"normalized": false,
|
17 |
+
"rstrip": false,
|
18 |
+
"single_word": false,
|
19 |
+
"special": true
|
20 |
+
},
|
21 |
+
"2": {
|
22 |
+
"content": "</s>",
|
23 |
+
"lstrip": false,
|
24 |
+
"normalized": false,
|
25 |
+
"rstrip": false,
|
26 |
+
"single_word": false,
|
27 |
+
"special": true
|
28 |
+
}
|
29 |
+
},
|
30 |
+
"bos_token": "<s>",
|
31 |
+
"clean_up_tokenization_spaces": false,
|
32 |
+
"eos_token": "</s>",
|
33 |
+
"legacy": false,
|
34 |
+
"model_max_length": 1000000000000000019884624838656,
|
35 |
+
"pad_token": "</s>",
|
36 |
+
"padding_side": "right",
|
37 |
+
"sp_model_kwargs": {},
|
38 |
+
"spaces_between_special_tokens": false,
|
39 |
+
"tokenizer_class": "LlamaTokenizer",
|
40 |
+
"trust_remote_code": false,
|
41 |
+
"unk_token": "<unk>",
|
42 |
+
"use_default_system_prompt": true,
|
43 |
+
"use_fast": true
|
44 |
+
}
|