Create README.md
Browse files
README.md
CHANGED
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
License: apache-2.0
|
3 |
+
Language:
|
4 |
+
- En
|
5 |
+
Pipeline_tag: text-generation
|
6 |
+
Base_model: IntervitensInc/Llama-3.1-Minitron-4B-Width-Base-chatml
|
7 |
+
tags:
|
8 |
+
- Chat
|
9 |
+
---
|
10 |
+
|
11 |
+
## This repo contains EXL2 quants of the model. If you need the original weights, please find them [here](https://huggingface.co/anthracite-org/magnum-v2-4b).
|
12 |
+
## Base repo only contains the measurement file, see revisions for your quant of choice.
|
13 |
+
|
14 |
+
- [measurement.json](https://huggingface.co/anthracite-org/magnum-v2-4b-exl2/tree/main)
|
15 |
+
- [2.0bpw](https://huggingface.co/anthracite-org/magnum-v2-4b-exl2/tree/2.0bpw)
|
16 |
+
- [3.0bpw](https://huggingface.co/anthracite-org/magnum-v2-4b-exl2/tree/3.0bpw)
|
17 |
+
- [4.0bpw](https://huggingface.co/anthracite-org/magnum-v2-4b-exl2/tree/4.0bpw)
|
18 |
+
- [5.0bpw](https://huggingface.co/anthracite-org/magnum-v2-4b-exl2/tree/5.0bpw)
|
19 |
+
- [6.0bpw](https://huggingface.co/anthracite-org/magnum-v2-4b-exl2/tree/6.0bpw)
|
20 |
+
- [8.0bpw](https://huggingface.co/anthracite-org/magnum-v2-4b-exl2/tree/8.0bpw)
|
21 |
+
|
22 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/658a46cbfb9c2bdfae75b3a6/9JwXZze4tHRGpc_RzE2AU.png)
|
23 |
+
This is the eighth in a series of models designed to replicate the prose quality of the Claude 3 models, specifically Sonnet and Opus. This model is fine-tuned on top of [IntervitensInc/Llama-3.1-Minitron-4B-Width-Base-chatml](https://huggingface.co/IntervitensInc/Llama-3.1-Minitron-4B-Width-Base-chatml).
|
24 |
+
|
25 |
+
## Prompting
|
26 |
+
Model has been Instruct tuned with the ChatML formatting. A typical input would look like this:
|
27 |
+
|
28 |
+
```py
|
29 |
+
"""<|im_start|>system
|
30 |
+
system prompt<|im_end|>
|
31 |
+
<|im_start|>user
|
32 |
+
Hi there!<|im_end|>
|
33 |
+
<|im_start|>assistant
|
34 |
+
Nice to meet you!<|im_end|>
|
35 |
+
<|im_start|>user
|
36 |
+
Can I ask a question?<|im_end|>
|
37 |
+
<|im_start|>assistant
|
38 |
+
"""
|
39 |
+
```
|
40 |
+
|
41 |
+
## Support
|
42 |
+
|
43 |
+
In order to inference this model you will have to use Aphrodite or vLLM as llama.cpp has not yet merged the required pull request to fix llama3.1 rope_freqs not respecting custom head_dim - You can however get around this by quanting the model yourself with the following fixes for a working GGUF. However, it will be stuck at 8k context until [this PR](https://github.com/ggerganov/llama.cpp/pull/9141) is merged.
|
44 |
+
|
45 |
+
1. Remove `"rope_scaling": {}` from `config.json`
|
46 |
+
2. Change `"max_position_embeddings"` to `8192` in `config.json`
|
47 |
+
3. Add `"add_bos_token": false` to `tokenizer_config.json`
|
48 |
+
|
49 |
+
|
50 |
+
## Credits
|
51 |
+
|
52 |
+
- [anthracite-org/Stheno-Data-Filtered](https://huggingface.co/datasets/anthracite-org/Stheno-Data-Filtered)
|
53 |
+
- [anthracite-org/kalo-opus-instruct-22k-no-refusal](https://huggingface.co/datasets/anthracite-org/kalo-opus-instruct-22k-no-refusal)
|
54 |
+
- [lodrick-the-lafted/NopmWritingStruct](https://huggingface.co/datasets/lodrick-the-lafted/NopmWritingStruct)
|
55 |
+
- [NewEden/Gryphe-3.5-16k-Subset](NewEden/Gryphe-3.5-16k-Subset)
|
56 |
+
- [Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned](https://huggingface.co/datasets/Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned)
|
57 |
+
- [Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned](https://huggingface.co/datasets/Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned)
|
58 |
+
|
59 |
+
This model has been a team effort, and the credits goes to all members of Anthracite.
|
60 |
+
|
61 |
+
## Training
|
62 |
+
The training was done for 2 epochs. We used 2 x [RTX 6000s](https://store.nvidia.com/en-us/nvidia-rtx/products/nvidia-rtx-6000-ada-generation/) GPUs graciously provided by [Kubernetes_Bad](https://huggingface.co/kubernetes-bad) for the full-parameter fine-tuning of the model.
|
63 |
+
|
64 |
+
[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
|
65 |
+
|
66 |
+
## Safety
|
67 |
+
...
|