Update README.md
Browse files
README.md
CHANGED
@@ -8,7 +8,28 @@ widget:
|
|
8 |
pipeline_tag: text-generation
|
9 |
---
|
10 |
|
11 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
12 |
|
13 |
#### Esiampl
|
14 |
|
@@ -101,4 +122,9 @@ while True:
|
|
101 |
print("Cymraeg:")
|
102 |
response = generate_text(prompt, settings, max_new_tokens)
|
103 |
print("="*132)
|
104 |
-
```
|
|
|
|
|
|
|
|
|
|
|
|
8 |
pipeline_tag: text-generation
|
9 |
---
|
10 |
|
11 |
+
# ALMA-Cymraeg-13B
|
12 |
+
Fersiwn Gymraeg o fodel cyfieithu [ALMA](https://github.com/fe1ixxu/ALMA) a ddisgrifir yn [https://arxiv.org/abs/2309.11674](https://arxiv.org/abs/2309.11674). \
|
13 |
+
_This is a Welsh version of the [ALMA](https://github.com/fe1ixxu/ALMA) LLM-based translation model._
|
14 |
+
|
15 |
+
Mae'r model LLM yn seiliedig ar Lama-2-13B, gyda hyfforddiant parhaus ar ddata Gymreig [OSCAR-2301](https://huggingface.co/datasets/oscar-corpus/OSCAR-2301) am 3 Epoch
|
16 |
+
ac yna hyfforddiant cywrain pellach ar ddata Cofnod y Cynulliad a ddarparir gan [TechIaith](https://huggingface.co/datasets/techiaith/cofnodycynulliad_en-cy).
|
17 |
+
|
18 |
+
Mae'r fersiwn yma wedi ei gywasgu i 4.0bpw er mwyn llwytho mewn 10GB o gof GPU wrth ddefnyddio [ExLlamaV2](https://github.com/turboderp/exllamav2).
|
19 |
+
|
20 |
+
### Fformat Sgwrs
|
21 |
+
|
22 |
+
Mae'r hyfforddiant cywrain wedi defnyddio'r fformat canlynol ar gyfer trosi o'r Saesneg i'r Gymraeg (a'r naill forrdd i'r llall).
|
23 |
+
```
|
24 |
+
Cyfieithwch y testun Saesneg canlynol i'r Gymraeg.
|
25 |
+
### Saesneg:
|
26 |
+
{prompt}
|
27 |
+
|
28 |
+
### Cymraeg:
|
29 |
+
|
30 |
+
```
|
31 |
+
|
32 |
+
|
33 |
|
34 |
#### Esiampl
|
35 |
|
|
|
122 |
print("Cymraeg:")
|
123 |
response = generate_text(prompt, settings, max_new_tokens)
|
124 |
print("="*132)
|
125 |
+
```
|
126 |
+
|
127 |
+
## Hawlfraint
|
128 |
+
|
129 |
+
Mae'r model yn seiliedig ar Llama2 ac felly dan drwydded gan [Meta](https://ai.meta.com/llama/license/). \
|
130 |
+
Mae'r data Cofnod y Cynulliad dan drywdded [Llywodraeth Agored](https://www.nationalarchives.gov.uk/doc/open-government-licence-cymraeg/version/3/).
|