RichardErkhov
commited on
Commit
โข
af794b5
1
Parent(s):
a7a6069
uploaded readme
Browse files
README.md
ADDED
@@ -0,0 +1,119 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
Quantization made by Richard Erkhov.
|
2 |
+
|
3 |
+
[Github](https://github.com/RichardErkhov)
|
4 |
+
|
5 |
+
[Discord](https://discord.gg/pvy7H8DZMG)
|
6 |
+
|
7 |
+
[Request more models](https://github.com/RichardErkhov/quant_request)
|
8 |
+
|
9 |
+
|
10 |
+
Hebrew-Gemma-11B-Instruct - GGUF
|
11 |
+
- Model creator: https://huggingface.co/yam-peleg/
|
12 |
+
- Original model: https://huggingface.co/yam-peleg/Hebrew-Gemma-11B-Instruct/
|
13 |
+
|
14 |
+
|
15 |
+
| Name | Quant method | Size |
|
16 |
+
| ---- | ---- | ---- |
|
17 |
+
| [Hebrew-Gemma-11B-Instruct.Q2_K.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q2_K.gguf) | Q2_K | 3.9GB |
|
18 |
+
| [Hebrew-Gemma-11B-Instruct.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.IQ3_XS.gguf) | IQ3_XS | 4.27GB |
|
19 |
+
| [Hebrew-Gemma-11B-Instruct.IQ3_S.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.IQ3_S.gguf) | IQ3_S | 4.48GB |
|
20 |
+
| [Hebrew-Gemma-11B-Instruct.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q3_K_S.gguf) | Q3_K_S | 4.48GB |
|
21 |
+
| [Hebrew-Gemma-11B-Instruct.IQ3_M.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.IQ3_M.gguf) | IQ3_M | 4.63GB |
|
22 |
+
| [Hebrew-Gemma-11B-Instruct.Q3_K.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q3_K.gguf) | Q3_K | 4.94GB |
|
23 |
+
| [Hebrew-Gemma-11B-Instruct.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q3_K_M.gguf) | Q3_K_M | 4.94GB |
|
24 |
+
| [Hebrew-Gemma-11B-Instruct.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q3_K_L.gguf) | Q3_K_L | 5.33GB |
|
25 |
+
| [Hebrew-Gemma-11B-Instruct.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.IQ4_XS.gguf) | IQ4_XS | 5.44GB |
|
26 |
+
| [Hebrew-Gemma-11B-Instruct.Q4_0.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q4_0.gguf) | Q4_0 | 5.68GB |
|
27 |
+
| [Hebrew-Gemma-11B-Instruct.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.IQ4_NL.gguf) | IQ4_NL | 5.72GB |
|
28 |
+
| [Hebrew-Gemma-11B-Instruct.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q4_K_S.gguf) | Q4_K_S | 5.72GB |
|
29 |
+
| [Hebrew-Gemma-11B-Instruct.Q4_K.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q4_K.gguf) | Q4_K | 6.04GB |
|
30 |
+
| [Hebrew-Gemma-11B-Instruct.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q4_K_M.gguf) | Q4_K_M | 6.04GB |
|
31 |
+
| [Hebrew-Gemma-11B-Instruct.Q4_1.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q4_1.gguf) | Q4_1 | 6.25GB |
|
32 |
+
| [Hebrew-Gemma-11B-Instruct.Q5_0.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q5_0.gguf) | Q5_0 | 6.81GB |
|
33 |
+
| [Hebrew-Gemma-11B-Instruct.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q5_K_S.gguf) | Q5_K_S | 6.81GB |
|
34 |
+
| [Hebrew-Gemma-11B-Instruct.Q5_K.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q5_K.gguf) | Q5_K | 7.0GB |
|
35 |
+
| [Hebrew-Gemma-11B-Instruct.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q5_K_M.gguf) | Q5_K_M | 7.0GB |
|
36 |
+
| [Hebrew-Gemma-11B-Instruct.Q5_1.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q5_1.gguf) | Q5_1 | 7.37GB |
|
37 |
+
| [Hebrew-Gemma-11B-Instruct.Q6_K.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q6_K.gguf) | Q6_K | 8.01GB |
|
38 |
+
| [Hebrew-Gemma-11B-Instruct.Q8_0.gguf](https://huggingface.co/RichardErkhov/yam-peleg_-_Hebrew-Gemma-11B-Instruct-gguf/blob/main/Hebrew-Gemma-11B-Instruct.Q8_0.gguf) | Q8_0 | 10.37GB |
|
39 |
+
|
40 |
+
|
41 |
+
|
42 |
+
|
43 |
+
Original model description:
|
44 |
+
---
|
45 |
+
license: other
|
46 |
+
license_name: gemma-terms-of-use
|
47 |
+
license_link: https://ai.google.dev/gemma/terms
|
48 |
+
language:
|
49 |
+
- en
|
50 |
+
- he
|
51 |
+
library_name: transformers
|
52 |
+
---
|
53 |
+
# Hebrew-Gemma-11B-Instruct
|
54 |
+
|
55 |
+
### Base Models:
|
56 |
+
- **07.03.2024:** [Hebrew-Gemma-11B](https://huggingface.co/yam-peleg/Hebrew-Gemma-11B)
|
57 |
+
- **16.03.2024:** [Hebrew-Gemma-11B-V2](https://huggingface.co/yam-peleg/Hebrew-Gemma-11B-V2)
|
58 |
+
|
59 |
+
### Instruct Models:
|
60 |
+
- **07.03.2024:** [Hebrew-Gemma-11B-Instruct](https://huggingface.co/yam-peleg/Hebrew-Gemma-11B-Instruct)
|
61 |
+
|
62 |
+
The Hebrew-Gemma-11B-Instruct Large Language Model (LLM) is a instruct fine-tuned version of the [Hebrew-Gemma-11B](https://huggingface.co/yam-peleg/Hebrew-Gemma-11B) generative text model using a variety of conversation datasets.
|
63 |
+
|
64 |
+
It is continued pretrain of gemma-7b, extended to a larger scale and trained on 3B additional tokens of both English and Hebrew text data.
|
65 |
+
|
66 |
+
|
67 |
+
# Instruction format
|
68 |
+
|
69 |
+
This format must be strictly respected, otherwise the model will generate sub-optimal outputs.
|
70 |
+
|
71 |
+
```
|
72 |
+
<bos><start_of_turn>user
|
73 |
+
Write a hello world program<end_of_turn>
|
74 |
+
<start_of_turn>model
|
75 |
+
Here is a simple hellow world program<end_of_turn><eos>
|
76 |
+
```
|
77 |
+
|
78 |
+
- The conversation starts with **`<bos>`**.
|
79 |
+
- Each turn is preceded by a **`<start_of_turn>`** delimiter and then the role of the entity (`user` or `model`).
|
80 |
+
- Turns finish with the **`<end_of_turn>`** token.
|
81 |
+
- Conversation finish with the **`<eos>`** token.
|
82 |
+
|
83 |
+
You can follow this format to build the prompt manually, if you need to do it without the tokenizer's chat template.
|
84 |
+
|
85 |
+
A simple example using the tokenizer's chat template:
|
86 |
+
|
87 |
+
```python
|
88 |
+
from transformers import AutoTokenizer, AutoModelForCausalLM
|
89 |
+
|
90 |
+
model_id = "Hebrew-Gemma-11B-Instruct"
|
91 |
+
|
92 |
+
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
93 |
+
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="cuda")
|
94 |
+
|
95 |
+
chat = [
|
96 |
+
{ "role": "user", "content": "ืืชืื ืงืื ืคืฉืื ืืคืืืชืื ืฉืืืคืืก ืืืกื ืืช ืืชืืจืื ืฉื ืืืื" },
|
97 |
+
]
|
98 |
+
prompt = tokenizer.apply_chat_template(chat, tokenize=False, add_generation_prompt=True)
|
99 |
+
```
|
100 |
+
|
101 |
+
### Terms of Use
|
102 |
+
|
103 |
+
As an extention of Gemma-7B, this model is subject to the original license and terms of use by Google.
|
104 |
+
|
105 |
+
### Benchmark Results
|
106 |
+
|
107 |
+
- Coming Soon!
|
108 |
+
|
109 |
+
|
110 |
+
### Notice
|
111 |
+
|
112 |
+
Hebrew-Gemma-11B is a pretrained base model and therefore does not have any moderation mechanisms.
|
113 |
+
|
114 |
+
|
115 |
+
### Authors
|
116 |
+
|
117 |
+
- Trained by Yam Peleg.
|
118 |
+
- In collaboration with Jonathan Rouach and Arjeo, inc.
|
119 |
+
|