Update README.md
Browse files
README.md
CHANGED
@@ -8,11 +8,11 @@ license: bigscience-openrail-m
|
|
8 |
---
|
9 |
|
10 |
|
11 |
-
[GeoV](https://
|
12 |
|
13 |
The GeoV model was designed by Georges Harik and uses
|
14 |
[Rotary Positional Embeddings with Relative distances (RoPER)](http://research.labml.ai/RoPER.html)
|
15 |
-
by [Georges
|
16 |
|
17 |
[RoPER]((http://research.labml.ai/RoPER.html),
|
18 |
in addition to using relative positions in the attention score calculation by RoPE embeddings,
|
@@ -43,25 +43,31 @@ The released weights were trained on ~70 billion tokens.
|
|
43 |
We plan to continue training up to 300 billion tokens and update the weights at every 20b tokens.
|
44 |
This training run is monolingual and uses c4en and english wikipedia datasets.
|
45 |
|
|
|
|
|
|
|
|
|
|
|
|
|
46 |
## Generation
|
47 |
|
48 |
-
|
49 |
|
50 |
```python
|
51 |
-
|
52 |
|
53 |
-
|
54 |
-
|
55 |
|
56 |
-
|
57 |
|
58 |
-
|
59 |
|
60 |
-
|
61 |
-
|
62 |
-
|
63 |
-
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
```
|
|
|
8 |
---
|
9 |
|
10 |
|
11 |
+
[GeoV](https://github.com/geov-ai/geov)-9B is a 9 billion parameter causal language model.
|
12 |
|
13 |
The GeoV model was designed by Georges Harik and uses
|
14 |
[Rotary Positional Embeddings with Relative distances (RoPER)](http://research.labml.ai/RoPER.html)
|
15 |
+
by [Georges Harik](https://twitter.com/gharik) and [Varuna Jayasiri](https://twitter.com/vpj).
|
16 |
|
17 |
[RoPER]((http://research.labml.ai/RoPER.html),
|
18 |
in addition to using relative positions in the attention score calculation by RoPE embeddings,
|
|
|
43 |
We plan to continue training up to 300 billion tokens and update the weights at every 20b tokens.
|
44 |
This training run is monolingual and uses c4en and english wikipedia datasets.
|
45 |
|
46 |
+
## Installation
|
47 |
+
|
48 |
+
```shell
|
49 |
+
pip install geov
|
50 |
+
```
|
51 |
+
|
52 |
## Generation
|
53 |
|
54 |
+
[](https://colab.research.google.com/github/geov-ai/geov/blob/master/notebooks/generate.ipynb)
|
55 |
|
56 |
```python
|
57 |
+
from geov import GeoVForCausalLM, GeoVTokenizer
|
58 |
|
59 |
+
model = GeoVForCausalLM.from_pretrained("GeoV/GeoV-9b")
|
60 |
+
tokenizer = GeoVTokenizer.from_pretrained("GeoV/GeoV-9b")
|
61 |
|
62 |
+
prompt = "In mathematics, topology is the study of"
|
63 |
|
64 |
+
input_ids = tokenizer(prompt, return_tensors="pt").input_ids
|
65 |
|
66 |
+
gen_tokens = model.generate(
|
67 |
+
input_ids,
|
68 |
+
do_sample=True,
|
69 |
+
temperature=0.9,
|
70 |
+
max_length=100,
|
71 |
+
)
|
72 |
+
gen_text = tokenizer.batch_decode(gen_tokens)[0]
|
73 |
```
|