Day 25: Open NLLB - filtering HBS (Pt 2)

preview_player
Показать описание
👨‍👩‍👧‍👦 Join our Discord community 👨‍👩‍👧‍👦

Continuing on with the data work (Serbian, Croatian, Bosnian) getting ready to train the first baseline English2HBS.

▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💰 BECOME A PATREON OF THE AI EPIPHANY ❤️

If these videos, GitHub projects, and blogs help you,
consider helping me out by supporting me on Patreon!

Huge thank you to these AI Epiphany patreons:
Eli Mahler
Petar Veličković

▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬

▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬

#opensource #meta #nolanguageleftbehind
  1. Aleksa Gordić - The AI Epiphany
  2. arxiv
  3. paper explained
  4. the ai epiphany
  5. ai
  6. deep learning
Рекомендации по теме
Day 24: Open 2:08:44

Day 24: Open NLLB - back from China, fuzzy dedup, preparing HBS run (Pt 1)

Day 22: Open 4:15:04

Day 22: Open NLLB - HBS data analysis, split into Cyrillic & Latin (Pt 1)

Day 10: Open 3:08:46

Day 10: Open NLLB - analyzing the primary dataset (Pt 2.)

Day 20: Open 2:03:56

Day 20: Open NLLB - analyzing HBS data (Bosnian, Croatian) (Pt 2)

Day 20: Open 2:52:57

Day 20: Open NLLB - downloading HBS parallel corpora (Croatian - wrap up) (Pt 1)

Day 15: Open 3:55:06

Day 15: Open NLLB - sharding, paper reading (hallucinations) (Pt 1)

Day 15: Open 2:34:06

Day 15: Open NLLB - wrapping up detecting hallucinations paper (Pt 3)

Day 18: Open 1:38:30

Day 18: Open NLLB - data loading document, GitHub tasks (Pt 1)

Day 10: Open 2:20:58

Day 10: Open NLLB - evaluation data, filtering (Pt 3.)

Day 19: Open 2:17:42

Day 19: Open NLLB - OPUS, downloading HBS parallel corpora (Pt 2)

Day 15: Open 1:12:54

Day 15: Open NLLB - ALTI+, detecting hallucinations (Pt 2)

Meta builds new 0:00:33

Meta builds new language translation model

NLLB-200, Video PreTraining, 1:03:30

NLLB-200, Video PreTraining, MineDojo... Trends in AI — July 2022

Day 1: Replicating 2:12:19

Day 1: Replicating Meta's NLLB - fairseq2, SeamlessM4T code, licenses (Pt. 3)

Morphological modeling in 1:54:39

Morphological modeling in Kinyarwanda Machine Translation and finetuning pretrained transformers.

Meta AI's new 0:10:47

Meta AI's new Translation AI | 200 Language Translation Model NLLB200 | Natural Language Proces...

Day 8: Meta 2:42:30

Day 8: Meta NLLB - analyzing filtered data, data preps (Pt. 2)

Redefining Domain Adaptation 0:58:12

Redefining Domain Adaptation for Machine Translation and Voice Recognition - An Essential Primer

[ML News] BLOOM: 0:14:03

[ML News] BLOOM: 176B Open-Source | Chinese Brain-Scale Computer | Meta AI: No Language Left Behind

# 175 Where 0:38:54

# 175 Where DeepL Beats ChatGPT with Graham Neubig

3% of Google 1:01:44

3% of Google code is by a LM, Data Pruning, Prompt Tuning... Trends in AI — August 2022

Toolformer LLM Can 0:23:35

Toolformer LLM Can Teach Themselves to Use API or Tools Paper Explanation Meta AI Research

Fine-tuning Large Language 0:28:18

Fine-tuning Large Language Models (LLMs) | w/ Example Code

The biggest data 0:12:08

The biggest data breach EVER: 1 BILLION people impacted