RichardErkhov
commited on
Commit
•
a173002
1
Parent(s):
f65f4ee
uploaded readme
Browse files
README.md
ADDED
@@ -0,0 +1,174 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
Quantization made by Richard Erkhov.
|
2 |
+
|
3 |
+
[Github](https://github.com/RichardErkhov)
|
4 |
+
|
5 |
+
[Discord](https://discord.gg/pvy7H8DZMG)
|
6 |
+
|
7 |
+
[Request more models](https://github.com/RichardErkhov/quant_request)
|
8 |
+
|
9 |
+
|
10 |
+
TinyLlama-1.1bee - GGUF
|
11 |
+
- Model creator: https://huggingface.co/BEE-spoke-data/
|
12 |
+
- Original model: https://huggingface.co/BEE-spoke-data/TinyLlama-1.1bee/
|
13 |
+
|
14 |
+
|
15 |
+
| Name | Quant method | Size |
|
16 |
+
| ---- | ---- | ---- |
|
17 |
+
| [TinyLlama-1.1bee.Q2_K.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q2_K.gguf) | Q2_K | 0.4GB |
|
18 |
+
| [TinyLlama-1.1bee.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.IQ3_XS.gguf) | IQ3_XS | 0.44GB |
|
19 |
+
| [TinyLlama-1.1bee.IQ3_S.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.IQ3_S.gguf) | IQ3_S | 0.47GB |
|
20 |
+
| [TinyLlama-1.1bee.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q3_K_S.gguf) | Q3_K_S | 0.47GB |
|
21 |
+
| [TinyLlama-1.1bee.IQ3_M.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.IQ3_M.gguf) | IQ3_M | 0.48GB |
|
22 |
+
| [TinyLlama-1.1bee.Q3_K.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q3_K.gguf) | Q3_K | 0.51GB |
|
23 |
+
| [TinyLlama-1.1bee.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q3_K_M.gguf) | Q3_K_M | 0.51GB |
|
24 |
+
| [TinyLlama-1.1bee.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q3_K_L.gguf) | Q3_K_L | 0.55GB |
|
25 |
+
| [TinyLlama-1.1bee.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.IQ4_XS.gguf) | IQ4_XS | 0.57GB |
|
26 |
+
| [TinyLlama-1.1bee.Q4_0.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q4_0.gguf) | Q4_0 | 0.59GB |
|
27 |
+
| [TinyLlama-1.1bee.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.IQ4_NL.gguf) | IQ4_NL | 0.6GB |
|
28 |
+
| [TinyLlama-1.1bee.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q4_K_S.gguf) | Q4_K_S | 0.6GB |
|
29 |
+
| [TinyLlama-1.1bee.Q4_K.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q4_K.gguf) | Q4_K | 0.62GB |
|
30 |
+
| [TinyLlama-1.1bee.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q4_K_M.gguf) | Q4_K_M | 0.62GB |
|
31 |
+
| [TinyLlama-1.1bee.Q4_1.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q4_1.gguf) | Q4_1 | 0.65GB |
|
32 |
+
| [TinyLlama-1.1bee.Q5_0.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q5_0.gguf) | Q5_0 | 0.71GB |
|
33 |
+
| [TinyLlama-1.1bee.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q5_K_S.gguf) | Q5_K_S | 0.71GB |
|
34 |
+
| [TinyLlama-1.1bee.Q5_K.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q5_K.gguf) | Q5_K | 0.73GB |
|
35 |
+
| [TinyLlama-1.1bee.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q5_K_M.gguf) | Q5_K_M | 0.73GB |
|
36 |
+
| [TinyLlama-1.1bee.Q5_1.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q5_1.gguf) | Q5_1 | 0.77GB |
|
37 |
+
| [TinyLlama-1.1bee.Q6_K.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q6_K.gguf) | Q6_K | 0.84GB |
|
38 |
+
| [TinyLlama-1.1bee.Q8_0.gguf](https://huggingface.co/RichardErkhov/BEE-spoke-data_-_TinyLlama-1.1bee-gguf/blob/main/TinyLlama-1.1bee.Q8_0.gguf) | Q8_0 | 1.09GB |
|
39 |
+
|
40 |
+
|
41 |
+
|
42 |
+
|
43 |
+
Original model description:
|
44 |
+
---
|
45 |
+
license: apache-2.0
|
46 |
+
base_model: PY007/TinyLlama-1.1B-intermediate-step-240k-503b
|
47 |
+
tags:
|
48 |
+
- bees
|
49 |
+
- beekeeping
|
50 |
+
- honey
|
51 |
+
metrics:
|
52 |
+
- accuracy
|
53 |
+
inference:
|
54 |
+
parameters:
|
55 |
+
max_new_tokens: 64
|
56 |
+
do_sample: true
|
57 |
+
renormalize_logits: true
|
58 |
+
repetition_penalty: 1.05
|
59 |
+
no_repeat_ngram_size: 6
|
60 |
+
temperature: 0.9
|
61 |
+
top_p: 0.95
|
62 |
+
epsilon_cutoff: 0.0008
|
63 |
+
widget:
|
64 |
+
- text: In beekeeping, the term "queen excluder" refers to
|
65 |
+
example_title: Queen Excluder
|
66 |
+
- text: One way to encourage a honey bee colony to produce more honey is by
|
67 |
+
example_title: Increasing Honey Production
|
68 |
+
- text: The lifecycle of a worker bee consists of several stages, starting with
|
69 |
+
example_title: Lifecycle of a Worker Bee
|
70 |
+
- text: Varroa destructor is a type of mite that
|
71 |
+
example_title: Varroa Destructor
|
72 |
+
- text: In the world of beekeeping, the acronym PPE stands for
|
73 |
+
example_title: Beekeeping PPE
|
74 |
+
- text: The term "robbing" in beekeeping refers to the act of
|
75 |
+
example_title: Robbing in Beekeeping
|
76 |
+
- text: |-
|
77 |
+
Question: What's the primary function of drone bees in a hive?
|
78 |
+
Answer:
|
79 |
+
example_title: Role of Drone Bees
|
80 |
+
- text: To harvest honey from a hive, beekeepers often use a device known as a
|
81 |
+
example_title: Honey Harvesting Device
|
82 |
+
- text: >-
|
83 |
+
Problem: You have a hive that produces 60 pounds of honey per year. You
|
84 |
+
decide to split the hive into two. Assuming each hive now produces at a 70%
|
85 |
+
rate compared to before, how much honey will you get from both hives next
|
86 |
+
year?
|
87 |
+
|
88 |
+
To calculate
|
89 |
+
example_title: Beekeeping Math Problem
|
90 |
+
- text: In beekeeping, "swarming" is the process where
|
91 |
+
example_title: Swarming
|
92 |
+
pipeline_tag: text-generation
|
93 |
+
datasets:
|
94 |
+
- BEE-spoke-data/bees-internal
|
95 |
+
language:
|
96 |
+
- en
|
97 |
+
---
|
98 |
+
|
99 |
+
|
100 |
+
# TinyLlama-1.1bee 🐝
|
101 |
+
|
102 |
+
|
103 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/60bccec062080d33f875cd0c/vgDfbjic0S3OJwv9BNzQN.png)
|
104 |
+
|
105 |
+
As we feverishly hit the refresh button on hf.co's homepage, on the hunt for the newest waifu chatbot to grace the AI stage, an epiphany struck us like a bee sting. What could we offer to the hive-mind of the community? The answer was as clear as honey—beekeeping, naturally. And thus, this un-bee-lievable model was born.
|
106 |
+
|
107 |
+
|
108 |
+
## Details
|
109 |
+
|
110 |
+
This model is a fine-tuned version of [PY007/TinyLlama-1.1B-intermediate-step-240k-503b](https://huggingface.co/PY007/TinyLlama-1.1B-intermediate-step-240k-503b) on the `BEE-spoke-data/bees-internal` dataset.
|
111 |
+
It achieves the following results on the evaluation set:
|
112 |
+
- Loss: 2.4285
|
113 |
+
- Accuracy: 0.4969
|
114 |
+
|
115 |
+
|
116 |
+
```
|
117 |
+
***** eval metrics *****
|
118 |
+
eval_accuracy = 0.4972
|
119 |
+
eval_loss = 2.4283
|
120 |
+
eval_runtime = 0:00:53.12
|
121 |
+
eval_samples = 239
|
122 |
+
eval_samples_per_second = 4.499
|
123 |
+
eval_steps_per_second = 1.129
|
124 |
+
perplexity = 11.3391
|
125 |
+
```
|
126 |
+
|
127 |
+
## 📜 Intended Uses & Limitations 📜
|
128 |
+
|
129 |
+
### Intended Uses:
|
130 |
+
|
131 |
+
1. **Educational Engagement**: Whether you're a novice beekeeper, an enthusiast, or someone just looking to understand the buzz around bees, this model aims to serve as an informative and entertaining resource.
|
132 |
+
2. **General Queries**: Have questions about hive management, bee species, or honey extraction? Feel free to consult the model for general insights.
|
133 |
+
3. **Academic & Research Inspiration**: If you're diving into the world of apiculture studies or environmental science, our model could offer some preliminary insights and ideas.
|
134 |
+
|
135 |
+
### Limitations:
|
136 |
+
|
137 |
+
1. **Not a Beekeeping Expert**: As much as we admire bees and their hard work, this model is not a certified apiculturist. Please consult professional beekeeping resources or experts for serious decisions related to hive management, bee health, and honey production.
|
138 |
+
2. **Licensing**: Apache-2.0, following TinyLlama
|
139 |
+
3. **Infallibility**: Our model can err, just like any other piece of technology (or bee). Always double-check the information before applying it to your own hive or research.
|
140 |
+
4. **Ethical Constraints**: This model may not be used for any illegal or unethical activities, including but not limited to: bioterrorism & standard terrorism, harassment, or spreading disinformation.
|
141 |
+
|
142 |
+
|
143 |
+
## Training and evaluation data
|
144 |
+
|
145 |
+
While the full dataset is not yet complete and therefore not yet released for "safety reasons", you can check out a preliminary sample at: [bees-v0](https://huggingface.co/datasets/BEE-spoke-data/bees-v0)
|
146 |
+
|
147 |
+
### Training hyperparameters
|
148 |
+
|
149 |
+
The following hyperparameters were used during training:
|
150 |
+
- learning_rate: 0.0001
|
151 |
+
- train_batch_size: 4
|
152 |
+
- eval_batch_size: 4
|
153 |
+
- seed: 80085
|
154 |
+
- gradient_accumulation_steps: 8
|
155 |
+
- total_train_batch_size: 32
|
156 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
157 |
+
- lr_scheduler_type: cosine
|
158 |
+
- lr_scheduler_warmup_ratio: 0.03
|
159 |
+
- num_epochs: 2.0
|
160 |
+
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
|
161 |
+
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_BEE-spoke-data__TinyLlama-1.1bee)
|
162 |
+
|
163 |
+
| Metric | Value |
|
164 |
+
|-----------------------|---------------------------|
|
165 |
+
| Avg. | 29.15 |
|
166 |
+
| ARC (25-shot) | 30.55 |
|
167 |
+
| HellaSwag (10-shot) | 51.8 |
|
168 |
+
| MMLU (5-shot) | 24.25 |
|
169 |
+
| TruthfulQA (0-shot) | 39.01 |
|
170 |
+
| Winogrande (5-shot) | 54.46 |
|
171 |
+
| GSM8K (5-shot) | 0.23 |
|
172 |
+
| DROP (3-shot) | 3.74 |
|
173 |
+
|
174 |
+
|