Total runs:
22.2K
Run Growth:
19.4K
Growth Rate:
87.18%
Sponsored by
APIMart
dicta-il / dictalm-7b-instruct
Model's Last Updated: November 06 2023
text-generation
Introduction of dictalm-7b-instruct
Model Details of dictalm-7b-instruct
DictaLM : A Large Generative Language Model for Modern Hebrew
A large generative pretrained transformer (GPT) language model for Hebrew, released here .
- This is an alpha version of the model, and there are many improvements to come.
- We are actively working on improving the model, so stay tuned.
This model was fine-tuned for instructions, here are a few examples of the different types of instructions the model was trained on:
-
General questions:
מה זה בית ספר?קיבלתי חתך קל באצבע. מהי הדרך הנכונה לטפל בזה? -
Simple tasks:
תציע כמה רעיונות לפעילות עם ילדים בני 5: -
Information retrieval from a paragraph context:
המסיק הידני הוא הדרך המסורתית והעתיקה לקטיף זיתים. שיטה זו דורשת כוח אדם רב באופן יחסי ועדיין מקובלת בישראל ובמקומות רבים בעולם. שיטות מסיק ידני מאפשרות חיסכון עלויות במקומות בהם כוח האדם זול ועלות השיטות הממוכנות גבוהה. לזיתים המיועדים למאכל (לכבישה, בניגוד לזיתים לשמן) מתאים יותר מסיק ידני כיוון שהפרי פחות נפגע במהלך המסיק בשיטה זו (פגיעות בקליפת הפרי בזיתים לשמן פחות משמעותיות). כמו כן מועדף מסיק ידני באזורים בהם הטופוגרפיה המקומית או צפיפות העצים לא מאפשרים גישה נוחה לכלים מכנים. השיטה הידנית מאפשרת גם למסוק עצים שונים במועדים שונים, בהתאם לקצב הבשלת הפרי הטבעי בכל עץ. על בסיס הפסקה הזאת, מה הוא היתרון של מסיק ידני מבחינת קצב הבשלת הפרי?
Sample usage:
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
tokenizer = AutoTokenizer.from_pretrained('dicta-il/dictalm-7b-instruct')
model = AutoModelForCausalLM.from_pretrained('dicta-il/dictalm-7b-instruct', trust_remote_code=True).cuda()
model.eval()
with torch.inference_mode():
prompt = 'תציע כמה רעיונות לפעילות עם ילדים בני 5:\n'
kwargs = dict(
inputs=tokenizer(prompt, return_tensors='pt').input_ids.to(model.device),
do_sample=True,
top_k=50,
top_p=0.95,
temperature=0.75,
max_length=100,
min_new_tokens=5
)
print(tokenizer.batch_decode(model.generate(**kwargs), skip_special_tokens=True))
There are many different parameters you can input into
kwargs
for different results (greedy, beamsearch, different sampling configurations, longer/shorter respones, etc.).
You can view the full list of parameters you can pass to the
generate
function
here
.
Alternative ways to initialize the model:
If you have multiple smaller GPUs, and the package
accelerate
is installed, you can initialize the model split across the devices:
model = AutoModelForCausalLM.from_pretrained('dicta-il/dictalm-7b-instruct', trust_remote_code=True, device_map='auto')
If you are running on linux and have the
bitsandbytes
package installed, you can initialize the model in 4/8 bit inference mode:
model = AutoModelForCausalLM.from_pretrained('dicta-il/dictalm-7b-instruct', trust_remote_code=True, load_in_8bit=True)
If you have FlashAttention installed in your environment, you can instruct the model to use the flash attention implementation (either V1 or V2, whichever is installed):
model = AutoModelForCausalLM.from_pretrained('dicta-il/dictalm-7b-instruct', trust_remote_code=True, use_flash_attention=True)
Colab notebook demos
You can try the model on a free tier google colab using the following notebooks:
- Streamlit based - you will need first to log in https://ngrok.com/ and get an authtoken, then paste it in the notebook ( screenshot ).
- Gradio based - uses deep speed for faster inference, text streamer to get the results as they are being generated and the UI is a widget embedded in the notebook ( screenshot ).
Citation
If you use DictaLM in your research, please cite
DictaLM -- A Large Generative Language Model for Modern Hebrew
BibTeX:
@misc{shmidman2023introducing,
title={Introducing DictaLM -- A Large Generative Language Model for Modern Hebrew},
author={Shaltiel Shmidman and Avi Shmidman and Amir David Nissan Cohen and Moshe Koppel},
year={2023},
eprint={2309.14568},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
License
Shield:
This work is licensed under a Creative Commons Attribution 4.0 International License .
Runs of dicta-il dictalm-7b-instruct on huggingface.co
134
Total runs
0
24-hour runs
0
3-day runs
0
7-day runs
0
30-day runs
More Information About dictalm-7b-instruct huggingface.co Model
Url of dictalm-7b-instruct
dictalm-7b-instruct huggingface.co Url
Provider of dictalm-7b-instruct huggingface.co
Other API from dicta-il
Updated:October 01 2026
Total runs:
14.4K
Run Growth:
-64
Growth Rate:
-0.44%
Total runs:
10.3K
Run Growth:
-1.0K
Growth Rate:
-9.79%
Total runs:
8.1K
Run Growth:
2.9K
Growth Rate:
36.51%
Total runs:
7.6K
Run Growth:
5.5K
Growth Rate:
73.04%
Total runs:
7.0K
Run Growth:
2.1K
Growth Rate:
30.12%
Total runs:
6.2K
Run Growth:
-7.7K
Growth Rate:
-122.66%
Total runs:
5.1K
Run Growth:
2.2K
Growth Rate:
43.46%
Total runs:
2.6K
Run Growth:
15
Growth Rate:
0.58%
Total runs:
2.2K
Run Growth:
1.3K
Growth Rate:
57.34%
Total runs:
2.2K
Run Growth:
559
Growth Rate:
25.71%
Total runs:
2.1K
Run Growth:
191
Growth Rate:
9.25%
Total runs:
1.5K
Run Growth:
709
Growth Rate:
46.40%
Total runs:
1.3K
Run Growth:
-1.5K
Growth Rate:
-113.91%
Total runs:
1.2K
Run Growth:
-2.9K
Growth Rate:
-237.17%
Total runs:
1.2K
Run Growth:
962
Growth Rate:
80.43%
Total runs:
695
Run Growth:
-333
Growth Rate:
-47.91%
Total runs:
666
Run Growth:
396
Growth Rate:
59.46%
Total runs:
522
Run Growth:
415
Growth Rate:
79.50%
Total runs:
408
Run Growth:
25
Growth Rate:
6.13%
Total runs:
362
Run Growth:
74
Growth Rate:
20.44%
Total runs:
362
Run Growth:
-282
Growth Rate:
-77.90%
Total runs:
334
Run Growth:
-850
Growth Rate:
-254.49%
Total runs:
323
Run Growth:
-720
Growth Rate:
-222.91%
Total runs:
323
Run Growth:
-9
Growth Rate:
-2.79%
Total runs:
309
Run Growth:
-150
Growth Rate:
-48.54%
Total runs:
235
Run Growth:
120
Growth Rate:
51.06%
Total runs:
215
Run Growth:
-274
Growth Rate:
-127.44%
Total runs:
155
Run Growth:
-190
Growth Rate:
-121.02%
Total runs:
141
Run Growth:
99
Growth Rate:
70.21%
Total runs:
113
Run Growth:
-23
Growth Rate:
-20.35%
Total runs:
71
Run Growth:
-451
Growth Rate:
-617.81%
Total runs:
59
Run Growth:
45
Growth Rate:
76.27%
Total runs:
56
Run Growth:
34
Growth Rate:
60.71%
Total runs:
48
Run Growth:
-9
Growth Rate:
-18.75%
Total runs:
33
Run Growth:
7
Growth Rate:
21.21%
Total runs:
33
Run Growth:
20
Growth Rate:
60.61%
Total runs:
30
Run Growth:
25
Growth Rate:
83.33%
Total runs:
30
Run Growth:
-92
Growth Rate:
-317.24%
Total runs:
28
Run Growth:
15
Growth Rate:
53.57%
Total runs:
22
Run Growth:
14
Growth Rate:
63.64%
Total runs:
21
Run Growth:
13
Growth Rate:
61.90%
Total runs:
20
Run Growth:
8
Growth Rate:
40.00%
Total runs:
19
Run Growth:
12
Growth Rate:
63.16%
Total runs:
18
Run Growth:
9
Growth Rate:
50.00%
Total runs:
17
Run Growth:
12
Growth Rate:
70.59%
Total runs:
17
Run Growth:
-21
Growth Rate:
-123.53%
Total runs:
16
Run Growth:
5
Growth Rate:
31.25%
Total runs:
12
Run Growth:
4
Growth Rate:
33.33%