afrideva commited on
Commit
627e53a
1 Parent(s): aa925fc

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +114 -0
README.md ADDED
@@ -0,0 +1,114 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: BEE-spoke-data/smol_llama-101M-GQA
3
+ datasets:
4
+ - JeanKaddour/minipile
5
+ - pszemraj/simple_wikipedia_LM
6
+ - BEE-spoke-data/wikipedia-20230901.en-deduped
7
+ - mattymchen/refinedweb-3m
8
+ inference: false
9
+ language:
10
+ - en
11
+ license: apache-2.0
12
+ model_creator: BEE-spoke-data
13
+ model_name: smol_llama-101M-GQA
14
+ pipeline_tag: text-generation
15
+ quantized_by: afrideva
16
+ tags:
17
+ - smol_llama
18
+ - llama2
19
+ - gguf
20
+ - ggml
21
+ - quantized
22
+ - q2_k
23
+ - q3_k_m
24
+ - q4_k_m
25
+ - q5_k_m
26
+ - q6_k
27
+ - q8_0
28
+ thumbnail: https://i.ibb.co/TvyMrRc/rsz-smol-llama-banner.png
29
+ widget:
30
+ - example_title: El Microondas
31
+ text: My name is El Microondas the Wise and
32
+ - example_title: Kennesaw State University
33
+ text: Kennesaw State University is a public
34
+ - example_title: Bungie
35
+ text: Bungie Studios is an American video game developer. They are most famous for
36
+ developing the award winning Halo series of video games. They also made Destiny.
37
+ The studio was founded
38
+ - example_title: Mona Lisa
39
+ text: The Mona Lisa is a world-renowned painting created by
40
+ - example_title: Harry Potter Series
41
+ text: The Harry Potter series, written by J.K. Rowling, begins with the book titled
42
+ - example_title: Riddle
43
+ text: 'Question: I have cities, but no houses. I have mountains, but no trees. I
44
+ have water, but no fish. What am I?
45
+
46
+ Answer:'
47
+ - example_title: Photosynthesis
48
+ text: The process of photosynthesis involves the conversion of
49
+ - example_title: Story Continuation
50
+ text: Jane went to the store to buy some groceries. She picked up apples, oranges,
51
+ and a loaf of bread. When she got home, she realized she forgot
52
+ - example_title: Math Problem
53
+ text: 'Problem 2: If a train leaves Station A at 9:00 AM and travels at 60 mph,
54
+ and another train leaves Station B at 10:00 AM and travels at 80 mph, when will
55
+ they meet if the distance between the stations is 300 miles?
56
+
57
+ To determine'
58
+ - example_title: Algorithm Definition
59
+ text: In the context of computer programming, an algorithm is
60
+ ---
61
+ # BEE-spoke-data/smol_llama-101M-GQA-GGUF
62
+
63
+ Quantized GGUF model files for [smol_llama-101M-GQA](https://huggingface.co/BEE-spoke-data/smol_llama-101M-GQA) from [BEE-spoke-data](https://huggingface.co/BEE-spoke-data)
64
+
65
+
66
+ | Name | Quant method | Size |
67
+ | ---- | ---- | ---- |
68
+ | [smol_llama-101m-gqa.fp16.gguf](https://huggingface.co/afrideva/smol_llama-101M-GQA-GGUF/resolve/main/smol_llama-101m-gqa.fp16.gguf) | fp16 | 203.28 MB |
69
+ | [smol_llama-101m-gqa.q2_k.gguf](https://huggingface.co/afrideva/smol_llama-101M-GQA-GGUF/resolve/main/smol_llama-101m-gqa.q2_k.gguf) | q2_k | 50.93 MB |
70
+ | [smol_llama-101m-gqa.q3_k_m.gguf](https://huggingface.co/afrideva/smol_llama-101M-GQA-GGUF/resolve/main/smol_llama-101m-gqa.q3_k_m.gguf) | q3_k_m | 57.06 MB |
71
+ | [smol_llama-101m-gqa.q4_k_m.gguf](https://huggingface.co/afrideva/smol_llama-101M-GQA-GGUF/resolve/main/smol_llama-101m-gqa.q4_k_m.gguf) | q4_k_m | 65.40 MB |
72
+ | [smol_llama-101m-gqa.q5_k_m.gguf](https://huggingface.co/afrideva/smol_llama-101M-GQA-GGUF/resolve/main/smol_llama-101m-gqa.q5_k_m.gguf) | q5_k_m | 74.34 MB |
73
+ | [smol_llama-101m-gqa.q6_k.gguf](https://huggingface.co/afrideva/smol_llama-101M-GQA-GGUF/resolve/main/smol_llama-101m-gqa.q6_k.gguf) | q6_k | 83.83 MB |
74
+ | [smol_llama-101m-gqa.q8_0.gguf](https://huggingface.co/afrideva/smol_llama-101M-GQA-GGUF/resolve/main/smol_llama-101m-gqa.q8_0.gguf) | q8_0 | 108.35 MB |
75
+
76
+
77
+
78
+ ## Original Model Card:
79
+ # smol_llama-101M-GQA
80
+
81
+ <img src="smol-llama-banner.png" alt="banner" style="max-width:95%; height:auto;">
82
+
83
+ A small 101M param (total) decoder model. This is the first version of the model.
84
+
85
+ - 768 hidden size, 6 layers
86
+ - GQA (24 heads, 8 key-value), context length 1024
87
+ - train-from-scratch
88
+
89
+ ## Notes
90
+
91
+ **This checkpoint** is the 'raw' pre-trained model and has not been tuned to a more specific task. **It should be fine-tuned** before use in most cases.
92
+
93
+ ### Checkpoints & Links
94
+
95
+ - _smol_-er 81M parameter checkpoint with in/out embeddings tied: [here](https://huggingface.co/BEE-spoke-data/smol_llama-81M-tied)
96
+ - Fine-tuned on `pypi` to generate Python code - [link](https://huggingface.co/BEE-spoke-data/smol_llama-101M-GQA-python)
97
+ - For the chat version of this model, please [see here](https://youtu.be/dQw4w9WgXcQ?si=3ePIqrY1dw94KMu4)
98
+
99
+ ---
100
+
101
+
102
+ # [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
103
+ Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_BEE-spoke-data__smol_llama-101M-GQA)
104
+
105
+ | Metric | Value |
106
+ |-----------------------|---------------------------|
107
+ | Avg. | 25.32 |
108
+ | ARC (25-shot) | 23.55 |
109
+ | HellaSwag (10-shot) | 28.77 |
110
+ | MMLU (5-shot) | 24.24 |
111
+ | TruthfulQA (0-shot) | 45.76 |
112
+ | Winogrande (5-shot) | 50.67 |
113
+ | GSM8K (5-shot) | 0.83 |
114
+ | DROP (3-shot) | 3.39 |