Upload README.md with huggingface_hub
Browse files
README.md
ADDED
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model: appvoid/no-prompt-1.3b
|
3 |
+
datasets:
|
4 |
+
- appvoid/no-prompt-15k
|
5 |
+
inference: false
|
6 |
+
language:
|
7 |
+
- en
|
8 |
+
license: apache-2.0
|
9 |
+
model_creator: appvoid
|
10 |
+
model_name: no-prompt-1.3b
|
11 |
+
pipeline_tag: text-generation
|
12 |
+
quantized_by: afrideva
|
13 |
+
tags:
|
14 |
+
- gguf
|
15 |
+
- ggml
|
16 |
+
- quantized
|
17 |
+
- q2_k
|
18 |
+
- q3_k_m
|
19 |
+
- q4_k_m
|
20 |
+
- q5_k_m
|
21 |
+
- q6_k
|
22 |
+
- q8_0
|
23 |
+
---
|
24 |
+
# appvoid/no-prompt-1.3b-GGUF
|
25 |
+
|
26 |
+
Quantized GGUF model files for [no-prompt-1.3b](https://huggingface.co/appvoid/no-prompt-1.3b) from [appvoid](https://huggingface.co/appvoid)
|
27 |
+
|
28 |
+
|
29 |
+
| Name | Quant method | Size |
|
30 |
+
| ---- | ---- | ---- |
|
31 |
+
| [no-prompt-1.3b.fp16.gguf](https://huggingface.co/afrideva/no-prompt-1.3b-GGUF/resolve/main/no-prompt-1.3b.fp16.gguf) | fp16 | 2.69 GB |
|
32 |
+
| [no-prompt-1.3b.q2_k.gguf](https://huggingface.co/afrideva/no-prompt-1.3b-GGUF/resolve/main/no-prompt-1.3b.q2_k.gguf) | q2_k | 631.52 MB |
|
33 |
+
| [no-prompt-1.3b.q3_k_m.gguf](https://huggingface.co/afrideva/no-prompt-1.3b-GGUF/resolve/main/no-prompt-1.3b.q3_k_m.gguf) | q3_k_m | 704.72 MB |
|
34 |
+
| [no-prompt-1.3b.q4_k_m.gguf](https://huggingface.co/afrideva/no-prompt-1.3b-GGUF/resolve/main/no-prompt-1.3b.q4_k_m.gguf) | q4_k_m | 873.27 MB |
|
35 |
+
| [no-prompt-1.3b.q5_k_m.gguf](https://huggingface.co/afrideva/no-prompt-1.3b-GGUF/resolve/main/no-prompt-1.3b.q5_k_m.gguf) | q5_k_m | 1.00 GB |
|
36 |
+
| [no-prompt-1.3b.q6_k.gguf](https://huggingface.co/afrideva/no-prompt-1.3b-GGUF/resolve/main/no-prompt-1.3b.q6_k.gguf) | q6_k | 1.17 GB |
|
37 |
+
| [no-prompt-1.3b.q8_0.gguf](https://huggingface.co/afrideva/no-prompt-1.3b-GGUF/resolve/main/no-prompt-1.3b.q8_0.gguf) | q8_0 | 1.43 GB |
|
38 |
+
|
39 |
+
|
40 |
+
|
41 |
+
## Original Model Card:
|
42 |
+
![palmer](https://huggingface.co/appvoid/no-prompt-1.3b/resolve/main/_ccd1a5dd-2ddc-4d5a-8163-fd6d1b39f5f4.jpeg?download=true)
|
43 |
+
# no-prompt
|
44 |
+
### a sheared-llama-1.3b fine-tuning
|
45 |
+
This model uses an 1.3 billion parameters model as base to be further fine-tuned on the same data as palmer. It works pretty good and even surpasses sota model on `hellaswag`.
|
46 |
+
|
47 |
+
### evaluation
|
48 |
+
|Model| ARC_C| HellaSwag| PIQA| Winogrande|
|
49 |
+
|------|-----|-----------|------|-------------|
|
50 |
+
|tinyllama-2t| 0.2807| 0.5463| 0.7067| 0.5683|
|
51 |
+
|palmer-001 | 0.2807| 0.5524| 0.7106| 0.5896|
|
52 |
+
|sheared-1.3b| 0.2910| 0.5935| 0.7339| 0.5809|
|
53 |
+
|no-prompt-1.3b| 0.3157| **0.6022**| 0.7334| 0.5864|
|
54 |
+
|falcon-rw-1b-instruct-openorca (sota) | **0.3362**| 0.5997| **0.7394**| **0.6148**|
|
55 |
+
|
56 |
+
This model was trained on less than 25% of the dataset yet achieves competitive performance to current sota on open llm leaderboard.
|
57 |
+
|
58 |
+
### training
|
59 |
+
Training took ~5 P100 gpu hours. It was trained on 15,000 gpt-4 shuffled samples. no-prompt was fine-tuned using lower learning rates ensuring it keeps as much general knowledge as possible.
|
60 |
+
|
61 |
+
### prompt
|
62 |
+
```
|
63 |
+
no prompt
|
64 |
+
```
|
65 |
+
|
66 |
+
### limitations
|
67 |
+
Hallucinations are frequent, just as any transformer model this size.
|
68 |
+
|
69 |
+
<a href="https://ko-fi.com/appvoid" target="_blank"><img src="https://cdn.buymeacoffee.com/buttons/v2/default-yellow.png" alt="Buy Me A Coffee" style="height: 48px !important;width: 180px !important; filter: invert(70%);" ></a>
|