Upload README.md
Browse files
README.md
CHANGED
@@ -1,10 +1,15 @@
|
|
1 |
---
|
|
|
|
|
|
|
2 |
inference: false
|
|
|
|
|
3 |
license: cc-by-nc-4.0
|
4 |
model_creator: AIDC-ai-business
|
5 |
-
model_link: https://huggingface.co/AIDC-ai-business/Marcoroni-7b
|
6 |
model_name: Marcoroni 7b
|
7 |
model_type: llama
|
|
|
8 |
quantized_by: TheBloke
|
9 |
---
|
10 |
|
@@ -62,11 +67,16 @@ Here is an incomplate list of clients and libraries that are known to support GG
|
|
62 |
<!-- repositories-available end -->
|
63 |
|
64 |
<!-- prompt-template start -->
|
65 |
-
## Prompt template:
|
66 |
|
67 |
```
|
|
|
|
|
|
|
68 |
{prompt}
|
69 |
|
|
|
|
|
70 |
```
|
71 |
|
72 |
<!-- prompt-template end -->
|
@@ -131,7 +141,7 @@ Refer to the Provided Files table below to see what files use which methods, and
|
|
131 |
Make sure you are using `llama.cpp` from commit [d0cee0d36d5be95a0d9088b674dbb27354107221](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
|
132 |
|
133 |
```shell
|
134 |
-
./main -ngl 32 -m marcoroni-7b.q4_K_M.gguf --color -c 4096 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "{prompt}"
|
135 |
```
|
136 |
|
137 |
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
|
@@ -222,5 +232,37 @@ And thank you again to a16z for their generous grant.
|
|
222 |
<!-- original-model-card start -->
|
223 |
# Original model card: AIDC-ai-business's Marcoroni 7b
|
224 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
225 |
|
226 |
<!-- original-model-card end -->
|
|
|
1 |
---
|
2 |
+
base_model: https://huggingface.co/AIDC-ai-business/Marcoroni-7b
|
3 |
+
datasets:
|
4 |
+
- Open-Orca/OpenOrca
|
5 |
inference: false
|
6 |
+
language:
|
7 |
+
- en
|
8 |
license: cc-by-nc-4.0
|
9 |
model_creator: AIDC-ai-business
|
|
|
10 |
model_name: Marcoroni 7b
|
11 |
model_type: llama
|
12 |
+
pipeline_tag: text-generation
|
13 |
quantized_by: TheBloke
|
14 |
---
|
15 |
|
|
|
67 |
<!-- repositories-available end -->
|
68 |
|
69 |
<!-- prompt-template start -->
|
70 |
+
## Prompt template: Alpaca
|
71 |
|
72 |
```
|
73 |
+
Below is an instruction that describes a task. Write a response that appropriately completes the request.
|
74 |
+
|
75 |
+
### Instruction:
|
76 |
{prompt}
|
77 |
|
78 |
+
### Response:
|
79 |
+
|
80 |
```
|
81 |
|
82 |
<!-- prompt-template end -->
|
|
|
141 |
Make sure you are using `llama.cpp` from commit [d0cee0d36d5be95a0d9088b674dbb27354107221](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
|
142 |
|
143 |
```shell
|
144 |
+
./main -ngl 32 -m marcoroni-7b.q4_K_M.gguf --color -c 4096 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "Below is an instruction that describes a task. Write a response that appropriately completes the request.\n\n### Instruction:\n{prompt}\n\n### Response:"
|
145 |
```
|
146 |
|
147 |
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
|
|
|
232 |
<!-- original-model-card start -->
|
233 |
# Original model card: AIDC-ai-business's Marcoroni 7b
|
234 |
|
235 |
+
# Marcoroni-7B
|
236 |
+
Fine-tuned from Llama2-7B,we use Orca-style data and other open source data for fine-tuning.
|
237 |
+
|
238 |
+
# Model Details
|
239 |
+
* **Trained by**: trained by AIDC AI-Business.
|
240 |
+
* **Model type:** **Marcoroni-7B** is an auto-regressive language model based on the Llama 2 transformer architecture.
|
241 |
+
* **Language(s)**: English
|
242 |
+
* **License for Marcoroni-7B base weights**: Non-Commercial Creative Commons license ([CC BY-NC-4.0](https://creativecommons.org/licenses/by-nc/4.0/))
|
243 |
+
|
244 |
+
|
245 |
+
# Prompting
|
246 |
+
|
247 |
+
## Prompt Template for alpaca style
|
248 |
+
|
249 |
+
```
|
250 |
+
### Instruction:
|
251 |
+
|
252 |
+
<prompt> (without the <>)
|
253 |
+
|
254 |
+
### Response:
|
255 |
+
```
|
256 |
+
|
257 |
+
|
258 |
+
# Evulation Results ([Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard))
|
259 |
+
|
260 |
+
| Metric | Value |
|
261 |
+
|-----------------------|-------|
|
262 |
+
| Avg. | 60.1 |
|
263 |
+
| ARC (25-shot) | 58.11 |
|
264 |
+
| HellaSwag (10-shot) | 80.08 |
|
265 |
+
| MMLU (5-shot) | 51.36 |
|
266 |
+
| TruthfulQA (0-shot) | 50.85 |
|
267 |
|
268 |
<!-- original-model-card end -->
|