|
--- |
|
license: cc-by-nc-4.0 |
|
language: |
|
- ro |
|
base_model: OpenLLM-Ro/RoMistral-7b-Instruct-2024-05-17 |
|
datasets: |
|
- OpenLLM-Ro/ro_sft_alpaca |
|
- OpenLLM-Ro/ro_sft_alpaca_gpt4 |
|
- OpenLLM-Ro/ro_sft_dolly |
|
- OpenLLM-Ro/ro_sft_selfinstruct_gpt4 |
|
- OpenLLM-Ro/ro_sft_norobots |
|
- OpenLLM-Ro/ro_sft_orca |
|
- OpenLLM-Ro/ro_sft_camel |
|
tags: |
|
- TensorBlock |
|
- GGUF |
|
model-index: |
|
- name: OpenLLM-Ro/RoMistral-7b-Instruct-2024-05-17 |
|
results: |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: RoMT-Bench |
|
type: RoMT-Bench |
|
metrics: |
|
- type: Score |
|
value: 4.99 |
|
name: Score |
|
- type: Score |
|
value: 5.46 |
|
name: First turn |
|
- type: Score |
|
value: 4.53 |
|
name: Second turn |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: RoCulturaBench |
|
type: RoCulturaBench |
|
metrics: |
|
- type: Score |
|
value: 3.38 |
|
name: Score |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: Romanian_Academic_Benchmarks |
|
type: Romanian_Academic_Benchmarks |
|
metrics: |
|
- type: accuracy |
|
value: 52.54 |
|
name: Average accuracy |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: OpenLLM-Ro/ro_arc_challenge |
|
type: OpenLLM-Ro/ro_arc_challenge |
|
metrics: |
|
- type: accuracy |
|
value: 50.41 |
|
name: Average accuracy |
|
- type: accuracy |
|
value: 47.47 |
|
name: 0-shot |
|
- type: accuracy |
|
value: 48.59 |
|
name: 1-shot |
|
- type: accuracy |
|
value: 50.3 |
|
name: 3-shot |
|
- type: accuracy |
|
value: 51.33 |
|
name: 5-shot |
|
- type: accuracy |
|
value: 52.36 |
|
name: 10-shot |
|
- type: accuracy |
|
value: 52.44 |
|
name: 25-shot |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: OpenLLM-Ro/ro_mmlu |
|
type: OpenLLM-Ro/ro_mmlu |
|
metrics: |
|
- type: accuracy |
|
value: 51.61 |
|
name: Average accuracy |
|
- type: accuracy |
|
value: 50.01 |
|
name: 0-shot |
|
- type: accuracy |
|
value: 50.18 |
|
name: 1-shot |
|
- type: accuracy |
|
value: 53.13 |
|
name: 3-shot |
|
- type: accuracy |
|
value: 53.12 |
|
name: 5-shot |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: OpenLLM-Ro/ro_winogrande |
|
type: OpenLLM-Ro/ro_winogrande |
|
metrics: |
|
- type: accuracy |
|
value: 66.48 |
|
name: Average accuracy |
|
- type: accuracy |
|
value: 64.96 |
|
name: 0-shot |
|
- type: accuracy |
|
value: 67.09 |
|
name: 1-shot |
|
- type: accuracy |
|
value: 67.01 |
|
name: 3-shot |
|
- type: accuracy |
|
value: 66.85 |
|
name: 5-shot |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: OpenLLM-Ro/ro_hellaswag |
|
type: OpenLLM-Ro/ro_hellaswag |
|
metrics: |
|
- type: accuracy |
|
value: 60.27 |
|
name: Average accuracy |
|
- type: accuracy |
|
value: 59.99 |
|
name: 0-shot |
|
- type: accuracy |
|
value: 59.48 |
|
name: 1-shot |
|
- type: accuracy |
|
value: 60.14 |
|
name: 3-shot |
|
- type: accuracy |
|
value: 60.61 |
|
name: 5-shot |
|
- type: accuracy |
|
value: 61.12 |
|
name: 10-shot |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: OpenLLM-Ro/ro_gsm8k |
|
type: OpenLLM-Ro/ro_gsm8k |
|
metrics: |
|
- type: accuracy |
|
value: 34.19 |
|
name: Average accuracy |
|
- type: accuracy |
|
value: 21.68 |
|
name: 1-shot |
|
- type: accuracy |
|
value: 38.21 |
|
name: 3-shot |
|
- type: accuracy |
|
value: 42.68 |
|
name: 5-shot |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: OpenLLM-Ro/ro_truthfulqa |
|
type: OpenLLM-Ro/ro_truthfulqa |
|
metrics: |
|
- type: accuracy |
|
value: 52.3 |
|
name: Average accuracy |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: LaRoSeDa_binary |
|
type: LaRoSeDa_binary |
|
metrics: |
|
- type: macro-f1 |
|
value: 97.36 |
|
name: Average macro-f1 |
|
- type: macro-f1 |
|
value: 97.27 |
|
name: 0-shot |
|
- type: macro-f1 |
|
value: 96.37 |
|
name: 1-shot |
|
- type: macro-f1 |
|
value: 97.97 |
|
name: 3-shot |
|
- type: macro-f1 |
|
value: 97.83 |
|
name: 5-shot |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: LaRoSeDa_multiclass |
|
type: LaRoSeDa_multiclass |
|
metrics: |
|
- type: macro-f1 |
|
value: 67.55 |
|
name: Average macro-f1 |
|
- type: macro-f1 |
|
value: 63.95 |
|
name: 0-shot |
|
- type: macro-f1 |
|
value: 66.89 |
|
name: 1-shot |
|
- type: macro-f1 |
|
value: 68.16 |
|
name: 3-shot |
|
- type: macro-f1 |
|
value: 71.19 |
|
name: 5-shot |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: LaRoSeDa_binary_finetuned |
|
type: LaRoSeDa_binary_finetuned |
|
metrics: |
|
- type: macro-f1 |
|
value: 98.8 |
|
name: Average macro-f1 |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: LaRoSeDa_multiclass_finetuned |
|
type: LaRoSeDa_multiclass_finetuned |
|
metrics: |
|
- type: macro-f1 |
|
value: 88.28 |
|
name: Average macro-f1 |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: WMT_EN-RO |
|
type: WMT_EN-RO |
|
metrics: |
|
- type: bleu |
|
value: 27.93 |
|
name: Average bleu |
|
- type: bleu |
|
value: 24.87 |
|
name: 0-shot |
|
- type: bleu |
|
value: 28.3 |
|
name: 1-shot |
|
- type: bleu |
|
value: 29.26 |
|
name: 3-shot |
|
- type: bleu |
|
value: 29.27 |
|
name: 5-shot |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: WMT_RO-EN |
|
type: WMT_RO-EN |
|
metrics: |
|
- type: bleu |
|
value: 13.21 |
|
name: Average bleu |
|
- type: bleu |
|
value: 3.69 |
|
name: 0-shot |
|
- type: bleu |
|
value: 5.45 |
|
name: 1-shot |
|
- type: bleu |
|
value: 19.92 |
|
name: 3-shot |
|
- type: bleu |
|
value: 23.8 |
|
name: 5-shot |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: WMT_EN-RO_finetuned |
|
type: WMT_EN-RO_finetuned |
|
metrics: |
|
- type: bleu |
|
value: 28.72 |
|
name: Average bleu |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: WMT_RO-EN_finetuned |
|
type: WMT_RO-EN_finetuned |
|
metrics: |
|
- type: bleu |
|
value: 40.86 |
|
name: Average bleu |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: XQuAD |
|
type: XQuAD |
|
metrics: |
|
- type: exact_match |
|
value: 43.66 |
|
name: Average exact_match |
|
- type: f1 |
|
value: 63.7 |
|
name: Average f1 |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: XQuAD_finetuned |
|
type: XQuAD_finetuned |
|
metrics: |
|
- type: exact_match |
|
value: 55.04 |
|
name: Average exact_match |
|
- type: f1 |
|
value: 72.31 |
|
name: Average f1 |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: STS |
|
type: STS |
|
metrics: |
|
- type: spearman |
|
value: 77.43 |
|
name: Average spearman |
|
- type: pearson |
|
value: 78.43 |
|
name: Average pearson |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: STS_finetuned |
|
type: STS_finetuned |
|
metrics: |
|
- type: spearman |
|
value: 87.25 |
|
name: Average spearman |
|
- type: pearson |
|
value: 87.79 |
|
name: Average pearson |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: XQuAD_EM |
|
type: XQuAD_EM |
|
metrics: |
|
- type: exact_match |
|
value: 23.36 |
|
name: 0-shot |
|
- type: exact_match |
|
value: 47.98 |
|
name: 1-shot |
|
- type: exact_match |
|
value: 51.85 |
|
name: 3-shot |
|
- type: exact_match |
|
value: 51.43 |
|
name: 5-shot |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: XQuAD_F1 |
|
type: XQuAD_F1 |
|
metrics: |
|
- type: f1 |
|
value: 46.29 |
|
name: 0-shot |
|
- type: f1 |
|
value: 67.4 |
|
name: 1-shot |
|
- type: f1 |
|
value: 70.58 |
|
name: 3-shot |
|
- type: f1 |
|
value: 70.53 |
|
name: 5-shot |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: STS_Spearman |
|
type: STS_Spearman |
|
metrics: |
|
- type: spearman |
|
value: 77.91 |
|
name: 1-shot |
|
- type: spearman |
|
value: 77.73 |
|
name: 3-shot |
|
- type: spearman |
|
value: 76.65 |
|
name: 5-shot |
|
- task: |
|
type: text-generation |
|
dataset: |
|
name: STS_Pearson |
|
type: STS_Pearson |
|
metrics: |
|
- type: pearson |
|
value: 78.03 |
|
name: 1-shot |
|
- type: pearson |
|
value: 78.74 |
|
name: 3-shot |
|
- type: pearson |
|
value: 78.53 |
|
name: 5-shot |
|
--- |
|
|
|
<div style="width: auto; margin-left: auto; margin-right: auto"> |
|
<img src="https://i.imgur.com/jC7kdl8.jpeg" alt="TensorBlock" style="width: 100%; min-width: 400px; display: block; margin: auto;"> |
|
</div> |
|
<div style="display: flex; justify-content: space-between; width: 100%;"> |
|
<div style="display: flex; flex-direction: column; align-items: flex-start;"> |
|
<p style="margin-top: 0.5em; margin-bottom: 0em;"> |
|
Feedback and support: TensorBlock's <a href="https://x.com/tensorblock_aoi">Twitter/X</a>, <a href="https://t.me/TensorBlock">Telegram Group</a> and <a href="https://x.com/tensorblock_aoi">Discord server</a> |
|
</p> |
|
</div> |
|
</div> |
|
|
|
## OpenLLM-Ro/RoMistral-7b-Instruct-2024-05-17 - GGUF |
|
|
|
This repo contains GGUF format model files for [OpenLLM-Ro/RoMistral-7b-Instruct-2024-05-17](https://huggingface.co/OpenLLM-Ro/RoMistral-7b-Instruct-2024-05-17). |
|
|
|
The files were quantized using machines provided by [TensorBlock](https://tensorblock.co/), and they are compatible with llama.cpp as of [commit b4242](https://github.com/ggerganov/llama.cpp/commit/a6744e43e80f4be6398fc7733a01642c846dce1d). |
|
|
|
<div style="text-align: left; margin: 20px 0;"> |
|
<a href="https://tensorblock.co/waitlist/client" style="display: inline-block; padding: 10px 20px; background-color: #007bff; color: white; text-decoration: none; border-radius: 5px; font-weight: bold;"> |
|
Run them on the TensorBlock client using your local machine ↗ |
|
</a> |
|
</div> |
|
|
|
## Prompt template |
|
|
|
``` |
|
<s>{system_prompt} [INST] {prompt} [/INST] |
|
``` |
|
|
|
## Model file specification |
|
|
|
| Filename | Quant type | File Size | Description | |
|
| -------- | ---------- | --------- | ----------- | |
|
| [RoMistral-7b-Instruct-2024-05-17-Q2_K.gguf](https://huggingface.co/tensorblock/RoMistral-7b-Instruct-2024-05-17-GGUF/blob/main/RoMistral-7b-Instruct-2024-05-17-Q2_K.gguf) | Q2_K | 2.719 GB | smallest, significant quality loss - not recommended for most purposes | |
|
| [RoMistral-7b-Instruct-2024-05-17-Q3_K_S.gguf](https://huggingface.co/tensorblock/RoMistral-7b-Instruct-2024-05-17-GGUF/blob/main/RoMistral-7b-Instruct-2024-05-17-Q3_K_S.gguf) | Q3_K_S | 3.165 GB | very small, high quality loss | |
|
| [RoMistral-7b-Instruct-2024-05-17-Q3_K_M.gguf](https://huggingface.co/tensorblock/RoMistral-7b-Instruct-2024-05-17-GGUF/blob/main/RoMistral-7b-Instruct-2024-05-17-Q3_K_M.gguf) | Q3_K_M | 3.519 GB | very small, high quality loss | |
|
| [RoMistral-7b-Instruct-2024-05-17-Q3_K_L.gguf](https://huggingface.co/tensorblock/RoMistral-7b-Instruct-2024-05-17-GGUF/blob/main/RoMistral-7b-Instruct-2024-05-17-Q3_K_L.gguf) | Q3_K_L | 3.822 GB | small, substantial quality loss | |
|
| [RoMistral-7b-Instruct-2024-05-17-Q4_0.gguf](https://huggingface.co/tensorblock/RoMistral-7b-Instruct-2024-05-17-GGUF/blob/main/RoMistral-7b-Instruct-2024-05-17-Q4_0.gguf) | Q4_0 | 4.109 GB | legacy; small, very high quality loss - prefer using Q3_K_M | |
|
| [RoMistral-7b-Instruct-2024-05-17-Q4_K_S.gguf](https://huggingface.co/tensorblock/RoMistral-7b-Instruct-2024-05-17-GGUF/blob/main/RoMistral-7b-Instruct-2024-05-17-Q4_K_S.gguf) | Q4_K_S | 4.140 GB | small, greater quality loss | |
|
| [RoMistral-7b-Instruct-2024-05-17-Q4_K_M.gguf](https://huggingface.co/tensorblock/RoMistral-7b-Instruct-2024-05-17-GGUF/blob/main/RoMistral-7b-Instruct-2024-05-17-Q4_K_M.gguf) | Q4_K_M | 4.368 GB | medium, balanced quality - recommended | |
|
| [RoMistral-7b-Instruct-2024-05-17-Q5_0.gguf](https://huggingface.co/tensorblock/RoMistral-7b-Instruct-2024-05-17-GGUF/blob/main/RoMistral-7b-Instruct-2024-05-17-Q5_0.gguf) | Q5_0 | 4.998 GB | legacy; medium, balanced quality - prefer using Q4_K_M | |
|
| [RoMistral-7b-Instruct-2024-05-17-Q5_K_S.gguf](https://huggingface.co/tensorblock/RoMistral-7b-Instruct-2024-05-17-GGUF/blob/main/RoMistral-7b-Instruct-2024-05-17-Q5_K_S.gguf) | Q5_K_S | 4.998 GB | large, low quality loss - recommended | |
|
| [RoMistral-7b-Instruct-2024-05-17-Q5_K_M.gguf](https://huggingface.co/tensorblock/RoMistral-7b-Instruct-2024-05-17-GGUF/blob/main/RoMistral-7b-Instruct-2024-05-17-Q5_K_M.gguf) | Q5_K_M | 5.131 GB | large, very low quality loss - recommended | |
|
| [RoMistral-7b-Instruct-2024-05-17-Q6_K.gguf](https://huggingface.co/tensorblock/RoMistral-7b-Instruct-2024-05-17-GGUF/blob/main/RoMistral-7b-Instruct-2024-05-17-Q6_K.gguf) | Q6_K | 5.942 GB | very large, extremely low quality loss | |
|
| [RoMistral-7b-Instruct-2024-05-17-Q8_0.gguf](https://huggingface.co/tensorblock/RoMistral-7b-Instruct-2024-05-17-GGUF/blob/main/RoMistral-7b-Instruct-2024-05-17-Q8_0.gguf) | Q8_0 | 7.696 GB | very large, extremely low quality loss - not recommended | |
|
|
|
|
|
## Downloading instruction |
|
|
|
### Command line |
|
|
|
Firstly, install Huggingface Client |
|
|
|
```shell |
|
pip install -U "huggingface_hub[cli]" |
|
``` |
|
|
|
Then, downoad the individual model file the a local directory |
|
|
|
```shell |
|
huggingface-cli download tensorblock/RoMistral-7b-Instruct-2024-05-17-GGUF --include "RoMistral-7b-Instruct-2024-05-17-Q2_K.gguf" --local-dir MY_LOCAL_DIR |
|
``` |
|
|
|
If you wanna download multiple model files with a pattern (e.g., `*Q4_K*gguf`), you can try: |
|
|
|
```shell |
|
huggingface-cli download tensorblock/RoMistral-7b-Instruct-2024-05-17-GGUF --local-dir MY_LOCAL_DIR --local-dir-use-symlinks False --include='*Q4_K*gguf' |
|
``` |
|
|
