cognitivelab_ai Profile Banner
CognitiveLab Profile
CognitiveLab

@cognitivelab_ai

Followers
1K
Following
24
Media
8
Statuses
24

Democratizing Generative AI

India
Joined July 2023
Don't wanna be here? Send us removal request.
@cognitivelab_ai
CognitiveLab
7 days
Also excited to release "M3DR: Towards Universal Multilingual Multimodal Document Retrieval" paper preprint that cover the framework we used to train these models
1
3
23
@cognitivelab_ai
CognitiveLab
7 days
Achieves SoTA performance on both cross lingual and monolingual benchmarks on our newly released NayanaIR-Bench
1
1
16
@cognitivelab_ai
CognitiveLab
7 days
Today we are excited to launch NetraEmbed SoTA multimodal multilingual document retrieval model. > Supports 22 languages > ~ 150% improvement over existing baselines > NayanaIR-Bench a open source multilingual document retrieval benchmark
5
28
193
@adithya_s_k
Adithya S K
2 months
Happy to share got a paper accepted at a neurips workshop This paper mainly tackles long context reasoning for multimodal models(vlms) with key focus citations and reasoning based re ranking got some good results with sft + rl(grpo) will put it on arxiv soon!
23
6
294
@cognitivelab_ai
CognitiveLab
8 months
Weโ€™re deeply honored by @Meta /@AIatMeta support and grateful to our incredible team for driving Nayana forward Stay tuned as we push the boundaries of multilingual, multimodal, multitask AI.
0
0
10
@cognitivelab_ai
CognitiveLab
8 months
Nayanaโ€™s Mission: Democratize AI for education, healthcare, governance, & cultural preservation across languages & modalities. Official Announcement : https://t.co/qRloKAZvhj
Tweet card summary image
about.fb.com
Weโ€™re excited to introduce the 10 international recipients of the second Llama Impact Grants.
1
0
7
@cognitivelab_ai
CognitiveLab
8 months
@cognitivelab_ai latest Milestones: ๐Ÿ“Œ Poster at #llamacon ๐Ÿ“Œ Papers accepted at NAACL & CVPR workshops ๐Ÿ“Œ Training on millions of synthetic datasets We're replacing fragmented AI pipelines with ONE cohesive model!
1
0
4
@cognitivelab_ai
CognitiveLab
8 months
Meet Nayana ("eyes" in Sanskrit): A unified AI model for text, vision, & audio! โœ… Supports 22 languages (10 Indic + 12 global) โœ… OCR, translation, Q&A, summarization many more tasks Know more :
Tweet card summary image
cognitivelab.in
CognitiveLab is proud to be selected as a recipient of Meta's prestigious Llama Impact Grant 2024, accelerating our revolutionary Nayana projectโ€”a multilingual (22 languages, 10 Indic), multimodal AI...
1
0
6
@cognitivelab_ai
CognitiveLab
8 months
๐Ÿš€ Thrilled to announce CognitiveLab wins the six-figure Llama Impact Grant by @Meta ! ๐Ÿ‡ฎ๐Ÿ‡ณ As India's only recipient, we're powering Nayana a multimodal, multilingual multi task AI model family! #AI #MultilingualAI #Innovation
2
0
34
@cognitivelab_ai
CognitiveLab
2 years
๐Ÿ‘‰๐Ÿผ Finetune for multi turn chat conversation ๐Ÿ‘‰๐Ÿผ Performers decent in cross lingual tasks ๐Ÿ‘‰๐Ÿผ Will soon put out the evals on Indic LLM leader board https://t.co/C2FWveiRro
Tweet card summary image
huggingface.co
0
0
3
@cognitivelab_ai
CognitiveLab
2 years
Our first attempt at fine-tuning LLama3 for Indic languages. we choose Hindi as LLama3 was efficient at tokenising it Many more experiments/iterations are in progress
1
0
1
@cognitivelab_ai
CognitiveLab
2 years
๐Ÿšจ Introducing Gaja ๐Ÿšจ ~ (Llama3-Gaja) A series of open source bilingual Hindi-English LLMs finetuned on top of Llama3-8b by @AIatMeta https://t.co/EJnjVgP387
1
8
75
@adithya_s_k
Adithya S K
2 years
evaluated LLama3 using indic_eval ( https://t.co/KFEXLwzxFM) on English, Hindi and Kannada results are there on the Indic LLM Leader board https://t.co/G8OfxTyCzV ๐Ÿ‘‰๐Ÿผ It performs significantly better than LLama2 in most of the benchmarks. ๐Ÿ‘‰๐Ÿผ In comparison to Gemma, it falls a
Tweet card summary image
huggingface.co
@adithya_s_k
Adithya S K
2 years
It's going to be hard to adapt Llama3 for Indic languages, in my opinion. Here are a few reasons why: ๐Ÿ‘‰๐Ÿผ The tokenizer used is TikToken-based, which is not really efficient in tokenizing Indic text despite having a vocabular size of 121k. ๐Ÿ‘‰๐Ÿผ unlike sentence-piece based models,
1
3
29
@cognitivelab_ai
CognitiveLab
2 years
The heart of the Leaderboard is indic_eval - https://t.co/kRcitPXXNo is an evaluation library that integrates seamlessly with the indic LLM leaderboard.
Tweet card summary image
github.com
A lightweight evaluation suite tailored specifically for assessing Indic LLMs across a diverse range of tasks - adithya-s-k/indic_eval
0
0
1
@cognitivelab_ai
CognitiveLab
2 years
Features include: ๐Ÿ‘‰๐Ÿผ Support for 7 Indic languages. ๐Ÿ‘‰๐Ÿผ Open source, hosted on Hugging Face. ๐Ÿ‘‰๐Ÿผ Support for 4 Indic benchmarks, with more to be added. ๐Ÿ‘‰๐Ÿผ Seamless integration with indic_eval This is the alpha release and we will be adding a lot more tested features soon
1
0
1
@cognitivelab_ai
CognitiveLab
2 years
Impressive Indic LLMs are emerging weekly, but without a standardized evaluation method or leaderboard, meaningful comparisons are difficult, hampering research and innovation. ๐Ÿ‘‰๐Ÿผ That's why we built - an open-source platform for comparing Indic LLMs across various benchmarks.
1
0
0
@cognitivelab_ai
CognitiveLab
2 years
Introducing ๐ŸšจIndic LLM Leaderboard๐Ÿšจ (alpha release)
1
5
17
@cognitivelab_ai
CognitiveLab
2 years
Here is the blog going into details about the model https://t.co/cG7LM6CX9X You can find the models here https://t.co/MCScJ6s0RS
Tweet card summary image
huggingface.co
0
0
7
@cognitivelab_ai
CognitiveLab
2 years
Our inaugural models, ๐—”๐—บ๐—ฏ๐—ฎ๐—ฟ๐—ถ-๐Ÿณ๐—•-๐—ฏ๐—ฎ๐˜€๐—ฒ-๐˜ƒ๐Ÿฌ.๐Ÿญ and ๐—”๐—บ๐—ฏ๐—ฎ๐—ฟ๐—ถ-๐Ÿณ๐—•-๐—œ๐—ป๐˜€๐˜๐—ฟ๐˜‚๐—ฐ๐˜-๐˜ƒ๐Ÿฌ.๐Ÿญ, achieve impressive results on a compact 1 billion-token training dataset, trained across multiple stages.
1
0
5