Update README.md
Browse files
README.md
CHANGED
|
@@ -1,51 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
-
|
| 8 |
-
-
|
| 9 |
-
|
| 10 |
-
- math
|
| 11 |
-
- factual
|
| 12 |
-
- agent
|
| 13 |
-
datasets:
|
| 14 |
-
- hotboxxgenn/mix-openhermes-openorca-platypus-airoboros-chatalpaca-opencode
|
| 15 |
-
- microsoft/rStar-Coder
|
| 16 |
-
- ed001/ds-coder-instruct-v1
|
| 17 |
-
- bigcode/starcoderdata
|
| 18 |
-
- bigcode/starcoder2data-extras
|
| 19 |
-
- codeparrot/self-instruct-starcoder
|
| 20 |
-
- mrtoy/mobile-ui-design
|
| 21 |
-
- YashJain/UI-Elements-Detection-Dataset
|
| 22 |
-
- tecky-tech/Tecky-UI-Elements-VLM
|
| 23 |
-
- Tesslate/UIGEN-T2
|
| 24 |
-
- FineWeb
|
| 25 |
-
- OpenWebMath
|
| 26 |
-
- UltraChat
|
| 27 |
-
- WizardCoderData
|
| 28 |
-
base_model:
|
| 29 |
-
- mistralai/Mixtral-8x7B-v0.1
|
| 30 |
-
- OpenBuddy/openbuddy-openllama-7b-v12-bf16
|
| 31 |
-
- HuggingFaceH4/mistral-7b-grok
|
| 32 |
-
- togethercomputer/RedPajama-INCITE-7B-Chat
|
| 33 |
---
|
| 34 |
|
| 35 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 36 |
|
| 37 |
-
|
| 38 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 39 |
|
| 40 |
---
|
| 41 |
|
| 42 |
-
##
|
| 43 |
-
|
| 44 |
-
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
|
| 48 |
|
| 49 |
---
|
| 50 |
|
| 51 |
-
##
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# ๐ซ BerryAI
|
| 2 |
+
|
| 3 |
+
BerryAI is a multimodal, coding-focused chatbot trained on a mixture of conversational, coding, UI design, and math reasoning datasets.
|
| 4 |
+
It blends creative dialogue with factual accuracy and programming skill.
|
| 5 |
+
|
| 6 |
---
|
| 7 |
+
|
| 8 |
+
## ๐ Features
|
| 9 |
+
- ๐ฌ Conversational: Fine-tuned with UltraChat for engaging responses
|
| 10 |
+
- ๐ง Smart: FineWeb improves general knowledge
|
| 11 |
+
- ๐งฎ Reasoning: OpenWebMath for step-by-step logic
|
| 12 |
+
- ๐ป Coder: WizardCoderData, StarCoderData, and DS-Coder-Instruct for deeper programming skill
|
| 13 |
+
- ๐จ UI Genius: Trained on multiple UI/UX datasets for design insights
|
| 14 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
---
|
| 16 |
|
| 17 |
+
## ๐งฐ Base Models
|
| 18 |
+
- `mistralai/Mixtral-8x7B-v0.1`
|
| 19 |
+
- `OpenBuddy/openbuddy-openllama-7b-v12-bf16`
|
| 20 |
+
- `HuggingFaceH4/mistral-7b-grok`
|
| 21 |
+
- `togethercomputer/RedPajama-INCITE-7B-Chat`
|
| 22 |
|
| 23 |
+
---
|
| 24 |
+
|
| 25 |
+
## ๐ Datasets
|
| 26 |
+
- hotboxxgenn/mix-openhermes-openorca-platypus-airoboros-chatalpaca-opencode
|
| 27 |
+
- microsoft/rStar-Coder
|
| 28 |
+
- ed001/ds-coder-instruct-v1
|
| 29 |
+
- bigcode/starcoderdata
|
| 30 |
+
- bigcode/starcoder2data-extras
|
| 31 |
+
- codeparrot/self-instruct-starcoder
|
| 32 |
+
- mrtoy/mobile-ui-design
|
| 33 |
+
- YashJain/UI-Elements-Detection-Dataset
|
| 34 |
+
- tecky-tech/Tecky-UI-Elements-VLM
|
| 35 |
+
- Tesslate/UIGEN-T2
|
| 36 |
+
- FineWeb
|
| 37 |
+
- OpenWebMath
|
| 38 |
+
- UltraChat
|
| 39 |
+
- WizardCoderData
|
| 40 |
|
| 41 |
---
|
| 42 |
|
| 43 |
+
## ๐ง Capabilities
|
| 44 |
+
โ
Coding (Python, C#, JavaScript)
|
| 45 |
+
โ
UI Design Advice (Mobile / Web)
|
| 46 |
+
โ
Conversational + Creative Writing
|
| 47 |
+
โ
Factual Knowledge
|
| 48 |
+
โ
Math & Logic Reasoning
|
| 49 |
|
| 50 |
---
|
| 51 |
|
| 52 |
+
## โก Usage
|
| 53 |
+
```python
|
| 54 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
| 55 |
+
|
| 56 |
+
model_id = "hotboxxgenn/BerryAI"
|
| 57 |
+
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
| 58 |
+
model = AutoModelForCausalLM.from_pretrained(model_id)
|
| 59 |
+
|
| 60 |
+
prompt = "Write a C# function that scrapes titles from a webpage."
|
| 61 |
+
inputs = tokenizer(prompt, return_tensors="pt")
|
| 62 |
+
outputs = model.generate(**inputs, max_new_tokens=250)
|
| 63 |
+
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
|