Applied Scientist (Speech Translation) at @apptek_mclean | Previously PhD at UPV (@mllpresearch group). ASR, MT & NLP.
Valencia
Joined August 2019
- Tweets94
- Following110
- Followers65
- Likes222
Pinned Tweet
Presenting our #ICASSP2020 paper, "EUROPARL-ST: A MULTILINGUAL CORPUS FOR SPEECH TRANSLATION OF PARLIAMENTARY DEBATES".
New resource for Speech Translation. 6 languages (En, Es, De, Fr, Pt, It) as both src and tgt.
arxiv.org/abs/1911.03167
2020.ieeeicassp-virtual.org/…
#nlp #NLProc
5
Javier Iranzo-Sánchez retweeted
Discovery of the day:
HF Transformers' gradient accumulation can significantly degrade the results compared to full batch training!
E.g.
I expected per_device_train_batch_size=1 and gradient_accumulation_steps=32 to be equivalent to per_device_train_batch_size=32 and gradient_accumulation_steps=1... but the former is much worse.
I ran several experiments with SmolLM-135M and Llama 3.2 1B. Using gradient accumulation consistently poorly performs for small batch sizes (same seed, single GPU).
Do I misunderstand something here?
I have no idea why gradient accumulation would have such a negative effect.
Check out our work at @aclmeeting, presented by my colleague @jorirsan. Drop by and have a chat about all topics streaming MT/ST.
Had a lot of fun today presenting our paper and talking with people at #ACL2024NLP :)
If you missed my oral presentation at the Machine Translation session, have some lingering questions or just want to chat for a bit, I will be tomorrow at 10:30h @ In-Person Poster Session 4 !
3
85
Javier Iranzo-Sánchez retweeted
Currently on my way to Bangkok to attend @aclmeeting #ACL2024 to represent our group @mllpresearch !!!
I will be presenting our TACL paper "Segmentation-Free Streaming Machine Translation" !
This is my first conference so I'm very excited to meet everyone in the community :)
I'm currently working as an Applied Scientist at @apptek_mclean, where I will keep working on exciting Speech and NLP projects. Looking forward to building great things together!
1
44
I really enjoyed my time at UPV, and I will always cherish the years working with my advisors and the rest of the @mllpresearch team, but it was time to move on from academia.
1
1
46
I defended my PhD thesis 2 weeks ago. Now that the dust has settled, I took some time to write down how everything turned out: jairsan.github.io/posts/2023…. I also recorded a mock PhD defense in case you are interested: youtube.com/watch?v=aqzrFUnC…
1
3
105
Feels like I just woke up from a strange dream. PhD successfully defended today. More updates soon.
9
210
Javier Iranzo-Sánchez retweeted
YYYY-MM-DD is obviously the best date format: it's the one where alphanumerical order matches chronological order.
What have I learned about developing ML research software during my PhD? My latest post: jairsan.github.io/posts/2023… discusses some of the critical issues I encountered, and how can you try to overcome them.
Congratulations to my dear friend @raruidol, who recently defended his PhD thesis. Ramon produced an impressive amount of research, and I have been lucky to collaborate with him in VivesDebate-Spech(arxiv.org/abs/2302.12584). Looking forward to hear about your next achievements!
👨🏻💻 Ramon Ruiz Dolz, #VRAIN member, has succesfully defended the following thesis about "Computational Argumentation for the Automatic Analysis of Argumentative Discourse and Human Persuasion".
⬇️ Check out in our website for more information!
vrain.upv.es/new-thesis-defe…
2
121
Javier Iranzo-Sánchez retweeted
It is admirable to apply the precautionary principle, and build & deploy transformative AI technology with exceptional care. But how is that best achieved by distracting from very real AI risks by making a remote, fanciful risk of extinction from AI a global priority? 🤔
Javier Iranzo-Sánchez retweeted
⚠️Did #ChatGPT cheat on your test? Probably yes
Many papers have evaluated ChatGPT on various benchmarks. However, it is important to consider that LLMs might have seen and memorized these datasets during pretraining.
Read our latest blog post: hitz-zentroa.github.io/lm-co…
🧵1/5
Javier Iranzo-Sánchez retweeted
🏆 Areg Sarvazyan, graduado en Ingeniería Informática de la @upv, logró ganar el mejor premio de trabajo final de grado por parte de la @acia_cat, como becario colaborador en #VRAIN.
👇 ¡Lee el artículo de abajo para aprender más sobre su proyecto!
upv.es/noticias-upv/noticia-…
I wholeheartedly agree. I recently carried out a major refactoring of one of my early research toolkits, and in the process I managed to remove a very significant amount of redundant or duplicate lines. One of the best feelings in software development.
1
172
Javier Iranzo-Sánchez retweeted
If you installed PyTorch-nightly on Linux between Dec. 25 and Dec. 30, uninstall it and torchtriton immediately and use the latest nightly binaries.
Read the security advisory here: pytorch.org/blog/compromised…
Javier Iranzo-Sánchez retweeted
The MLLP's Javier Jorge Cano @javierjorgecano has earned his PhD in Computer Science at @UPV, on the topic "#StreamingAutomaticSpeechRecognition with Hybrid Architectures and Deep Neural Network Models". Congratulations, Javi! 🎓 Read all about it: mllp.upv.es/mllp-member-javi…
How should you conduct Speech Translation Evaluation? My latest blog post provides a brief overview that goes beyond pre-segmented academic datasets. This could be specially interesting if you are a newcomer to the field.
jairsan.github.io/posts/2022…
Javier Iranzo-Sánchez retweeted
Tried new tutorial in my MSc class on Evaluation of AI: went through GPT3 prompting paper and discussed numerous evaluation problems (following ehudreiter.com/2021/08/09/ex…). Worked well, fact that there are serious flaws in such a prominent paper had impact on students.
I recently created a personal website in order to share insights into my research, as well as other topics of interest. You can access it at jairsan.github.io/,
as well as reading my motivation behind this move here: jairsan.github.io/posts/2012…