Update README.md
Browse files
README.md
CHANGED
|
@@ -1,4 +1,6 @@
|
|
| 1 |
---
|
|
|
|
|
|
|
| 2 |
library_name: transformers
|
| 3 |
tags:
|
| 4 |
- 4-bit
|
|
@@ -6,6 +8,22 @@ tags:
|
|
| 6 |
- text-generation
|
| 7 |
- autotrain_compatible
|
| 8 |
- endpoints_compatible
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
pipeline_tag: text-generation
|
| 10 |
inference: false
|
| 11 |
quantized_by: Suparious
|
|
@@ -15,7 +33,26 @@ quantized_by: Suparious
|
|
| 15 |
- Model creator: [cognitivecomputations](https://huggingface.co/cognitivecomputations)
|
| 16 |
- Original model: [dolphin-2.9-llama3-8b-1m](https://huggingface.co/cognitivecomputations/dolphin-2.9-llama3-8b-1m)
|
| 17 |
|
|
|
|
| 18 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
|
| 20 |
## How to use
|
| 21 |
|
|
|
|
| 1 |
---
|
| 2 |
+
license: other
|
| 3 |
+
base_model: meta-llama/Meta-Llama-3-8B
|
| 4 |
library_name: transformers
|
| 5 |
tags:
|
| 6 |
- 4-bit
|
|
|
|
| 8 |
- text-generation
|
| 9 |
- autotrain_compatible
|
| 10 |
- endpoints_compatible
|
| 11 |
+
- generated_from_trainer
|
| 12 |
+
- axolotl
|
| 13 |
+
model-index:
|
| 14 |
+
- name: out
|
| 15 |
+
results: []
|
| 16 |
+
datasets:
|
| 17 |
+
- cognitivecomputations/Dolphin-2.9
|
| 18 |
+
- teknium/OpenHermes-2.5
|
| 19 |
+
- m-a-p/CodeFeedback-Filtered-Instruction
|
| 20 |
+
- cognitivecomputations/dolphin-coder
|
| 21 |
+
- cognitivecomputations/samantha-data
|
| 22 |
+
- HuggingFaceH4/ultrachat_200k
|
| 23 |
+
- microsoft/orca-math-word-problems-200k
|
| 24 |
+
- abacusai/SystemChat-1.1
|
| 25 |
+
- Locutusque/function-calling-chatml
|
| 26 |
+
- internlm/Agent-FLAN
|
| 27 |
pipeline_tag: text-generation
|
| 28 |
inference: false
|
| 29 |
quantized_by: Suparious
|
|
|
|
| 33 |
- Model creator: [cognitivecomputations](https://huggingface.co/cognitivecomputations)
|
| 34 |
- Original model: [dolphin-2.9-llama3-8b-1m](https://huggingface.co/cognitivecomputations/dolphin-2.9-llama3-8b-1m)
|
| 35 |
|
| 36 |
+
<img src="https://cdn-uploads.huggingface.co/production/uploads/63111b2d88942700629f5771/ldkN1J0WIDQwU4vutGYiD.png" width="600" />
|
| 37 |
|
| 38 |
+
## Model Summary
|
| 39 |
+
|
| 40 |
+
Curated and trained by Eric Hartford, Lucas Atkins, and Fernando Fernandes, and Cognitive Computations
|
| 41 |
+
|
| 42 |
+
This version of Dolphin has a 1 million token context. I have applied `winglian/llama-3-1m-context-gradient-lora` - created by @gradientai and @winglian and sponsored by @CrusoeCloud
|
| 43 |
+
|
| 44 |
+
A bug has been found in the Dolphin 2.9 dataset in SystemConversations that causes the model to overly talk about the "SYSTEM MESSAGE". To counter this, we recommend you add a statement in the system message directing the model not to mention the system message. An example system message is "The assistant is named Dolphin. A helpful and friendly AI assistant, Dolphin avoids discussing the system message unless directly asked about it."
|
| 45 |
+
|
| 46 |
+
My appreciation for the sponsors of Dolphin 2.9:
|
| 47 |
+
- [Crusoe Cloud](https://crusoe.ai/) - provided excellent on-demand 10xL40S node
|
| 48 |
+
|
| 49 |
+
This model is based on Llama-3-8b, and is governed by [META LLAMA 3 COMMUNITY LICENSE AGREEMENT](LICENSE)
|
| 50 |
+
|
| 51 |
+
The base model has 8k context, and the full-weight fine-tuning was with 4k sequence length.
|
| 52 |
+
|
| 53 |
+
It took 2.5 days on 8x L40S provided by Crusoe Cloud
|
| 54 |
+
|
| 55 |
+
This model was trained FFT on all parameters, using ChatML prompt template format.
|
| 56 |
|
| 57 |
## How to use
|
| 58 |
|