Hugging Face's logo

Language and Translation Technology Team

university
https://lt3.ugent.be/
lt3ugent
lt3
Activity Feed Request to join this org

AI & ML interests

terminology; sentiment analysis; natural language processing; emotion detection; machine translation

Recent Activity

TomMoeras  updated a dataset about 2 months ago
LT3/nfr_bt_nmt_english-ukrainian
TomMoeras  updated a dataset about 2 months ago
LT3/nfr_bt_nmt_english-french
BramVanroy  authored a paper about 1 year ago
Language Resources for Dutch Large Language Modelling
View all activity

LT3's activity

BramVanroy 
posted an update 16 days ago
view post
Post
421
In the spirit of "Better late than never", I've finally written a brief overview paper for GEITje 7B Ultra. Initially released 10 months ago (oops), but still reaching around 1300 monthly downloads across the HF ecosystem (not including ollama).

GEITje 7B Ultra: A Conversational Model for Dutch (2412.04092)

While the paper discusses the model a little bit, I especially wanted to write about the datasets, which to this day seem an important asset for Dutch LLM training (SFT and preference tuning). We have a long way to go for Dutch, but publishing transparent and reproducible artefacts seems an important step to me, alongside having open discussions about data, bias, architectures.

In that spirit, thanks are in order for the creation of GEITje 7B Ultra and all related datasets:

- Michiel Buisman and UWV for providing the means to create the datasets
- Flemish Supercomputer Center (VSC) for the compute
- The Hugging Face Fellows and rest of the team for their discussions and insights
- The Dutch NLP community, notably @Rijgersberg for building the base GEITje model and the fruitful discussions we've had

More to come, step by step!

BramVanroy/geitje-7b-ultra-65c1ee010ad80fd1f6a8f208
TomMoeras 
updated 2 datasets about 2 months ago
BramVanroy 
posted an update 7 months ago
view post
Post
1678
The InstructGPT paper mentions that they insert 10% pretraining data during SFT, which they find improves the effect of PPO (IIUC). Has anyone else done later ablations on this? I've only seen the inverse suggested, mixing in SFT data during pretraining.
  • 2 replies
·
BramVanroy 
posted an update 7 months ago
view post
Post
2297
All my models seem to be plagued by infinite lists. When you ask a question that requires it to write a list, it most often keeps adding bullet points or enumeration. I am wondering whether this is a result of using chatty GPT-4 as DPO preferences. Any thoughts?
  • 1 reply
·
BramVanroy 
posted an update 8 months ago
view post
Post
2286
🥳 New license for datasets: Apache 2.0!

I have been struggling mentally for many months now with the OpenAI terms of use that indicate that their model outputs cannot be used to build "competing models". This leads to many questions:

- what is the definition of competing? Is it the same as "commercial"?
- since this is part of the terms of use between OpenAI and the API user, can a third party still use the generated dataset to build competing models?
- are such restrictions even legal in the first place?

Trying to "follow the rules" as much as possible despite wanting to be as open as possible, I kept releasing my datasets under non-commercial licenses (which are too restrictive anyhow - nothing should prevent you from using the data in non-LM commercial settings), just like models trained on these datasets. This has put me at a competitive disadvantage compared to creators who do not follow the same approach and release their data/models on apache 2.0 despite the OpenAI "restrictions". Moreover, I fear (https://twitter.com/BramVanroy/status/1780220420316164246) that my approach blocks adaptation of my data/models for (commercial) applications/integrations.

Thankfully @Rijgersberg noted that these OpenAI terms of use are NOT explicit in the Azure OpenAI API (https://twitter.com/E_Rijgersberg/status/1780308971762450725). Since my latest datasets were created via Azure, this comes as a relief. As far as I can tell after digging through Azure docs, this allows me to change all recent GPT4-generated datasets to apache 2.0! 🥳

- BramVanroy/ultrachat_200k_dutch
- BramVanroy/orca_dpo_pairs_dutch
- BramVanroy/ultra_feedback_dutch
- BramVanroy/ultra_feedback_dutch_cleaned
- BramVanroy/no_robots_dutch

I will have to mull over what I'll do for the older GPT3.5 datasets. What do you think that I should do?
·
BramVanroy 
posted an update 9 months ago
view post
Post
2439
🎈 LLM Benchmarks Update!

**tl;dr: do not depend on benchmark leaderboards to choose your "chatbot" model! (Especially for non-English languages.)**

First of all, I'm discontinuing the Open #Dutch #LLM Leaderboard (https://lnkd.in/eFnsaFR6). It will stay online for now, but I urge the use of the ScandEval leaderboard instead (https://scandeval.com/dutch-nlg/) by @saattrupdan . It contains more tasks, has better reproducibility and statistics (CI) and a flexible back-end library (scandeval) to run your own benchmarks with. As part of project "Leesplank" (with Michiel Buisman and Maarten Lens-FitzGerald) we recently added GPT-4-1106-preview scores to add a good "target" to the leaderboard.

An important note here is that benchmark leaderboards are not a golden truth. Especially evaluating generative models is hard. You run into issues like prompt engineering (and sensitivity of models to one or other prompt), structured output generation, and - quite simply - "how to automatically evaluate open-ended generation".

💡 Another important but under-discussed facet is the discrepancy between models' capability of understanding vs. generating *in different languages* (so the NLU part of NLG benchmarking). In other words: some of the listed models score really well on, e.g., MCQ benchmarks but are not suitable to use as DUTCH chat bots. Interestingly, some of these models seem to understand questions in Dutch and are able to pick the right answer (because they have good knowledge or reasoning skills), but generating fluent and grammatical Dutch is something else entirely! This is perhaps also true for humans: it's easier to sort-of grasp the meaning of a new language and answer with "Yes" or "No", but answering fluently in the language is much harder! Yet, your language production fluency does not necessarily say anything about your knowledge and reasoning skills.

Hopefully we can get a chat arena for Dutch some day - user feedback is the most powerful metric!
·
BramVanroy 
posted an update 9 months ago
view post
Post
2392
Does anyone have experience with finetuning Gemma? Even the 2B variant feels more memory heavy than mistral 7B. I know that its vocabulary is much larger (250k) but I'm a bit surprised that the max batch size that I can get in an A100 80GB is only 2 whereas I could fit 4 with mistral 7B - even though Gemma is much smaller except for the embedding layer. Both runs were using FA, same sequence length, same deepspeed zero 3 settings. Oh and yes I'm using the most recent hot fix of transformers that solves a memory issue with Gemma and others.

Any prior experience that you can share or suggestions to improve throughout?
  • 4 replies
·
BramVanroy 
posted an update 9 months ago
view post
Post
1725
🖴 The HPLT monolingual dataset has a new home!

After being in touch with HPLT folks, I've transfered the data to their org. That only makes sense. You can find it below.

HPLT/hplt_monolingual_v1_2
·
BramVanroy 
posted an update 10 months ago
view post
Post
🗄️ Massive data release on the HF Hub for 75 languages!

https://huggingface.co/datasets/BramVanroy/hplt_monolingual_v1_2

In December of last year, HPLT (https://hplt-project.org/) released version 1.2 of their dataset. It covers web-crawled data of 75 languages!, in the raw format as well as deduplicated and cleaned sections. In total, we're talking about over 40TB of data! This data was already accessible via their website but I figured the accessibility could be improved by an integration with Hugging Face tooling. 🤗 So I added the dataset here to the Hugging Face hub, enabing direct use in your conventional training pipelines for LLMs or other language technologies. The data will automatically be downloaded and optimised with just one line of code:

load_dataset("BramVanroy/hplt_mono_v1_2", "nl_cleaned")

Let's use this big blob of data to build something awesome in our languages! 🥳
  • 1 reply
·
BramVanroy 
posted an update 11 months ago
view post
Post
📣 DPO Dutch model release + datasets

After teasing for a while, I am finally releasing **GEITje 7B Ultra**, building upon the great GEITje 7B by @Rijgersberg . New contributions include: large new datasets for SFT (instruction/chat), two datasets for DPO training (i.e. RLAIF), and an SFT and DPO version of GEITje. The READMEs describe everything well (I hope), and I'll also share more info on social medias tomorrow.

For me this is a huge release, the datasets more so than the models. I'm especially pleased with UltraChat, which I created with the intent of having a diverse dataset - the model must be able to communicate with different types of users. So the user questions are created as if they were written by different personas, e.g. language learners, young children, experts, critics, etc. The focus with this is "building a good communication bot that is accessible and can handle different kinds of user input".

I wish I could find the time to also write a paper to get some "academic recognition" but that'll have to wait for now. I just want to bring it to the public so that others can play with it and use it to build new, cool stuff!

I hope that you can all appreciate the work. Let's build some cool stuff with it!

Models:
- Demo: https://huggingface.co/spaces/BramVanroy/GEITje-7B-ultra
- DPO Model: BramVanroy/GEITje-7B-ultra
- SFT model (not recommended): BramVanroy/GEITje-7B-ultra-sft

Datasets with GPT-4 turbo completions:
- No robots (~10k instructions): BramVanroy/no_robots_dutch
- UltraChat (~200k instructions): BramVanroy/ultrachat_200k_dutch
- UltraFeedback (DPO with GPT4+GEITje chat, ~50k): BramVanroy/ultra_feedback_dutch
- Orca DPO Pairs (DPO with GPT4+GEITje chat, ~10k): BramVanroy/orca_dpo_pairs_dutch
·
BramVanroy 
posted an update 11 months ago
view post
Post
🔎 DPO hyperparameter search update!

In my previous post (https://huggingface.co/posts/BramVanroy/633544255876795), I indicated how despite high reward accuracies and low losses, my model would sometimes just output repeating random tokens (/*****/). There were some useful brainstorms in that thread. I think the dataset is relatively easy for the model, leading it to quickly overfit when the beta is very small, which allows the model to step away further from its initially outputs.

So, I ran a hyperparameter search for learning rate (1e-7 v 5e-7), batch size (32, 64, 96, 128) and most importantly, beta (0.01, 0.1, 0.2, 0.5). You can have a look at the results for yourself here: https://wandb.ai/bramvanroy/dpo-geitje-ultra-hyperparams

Interpreting the result, I'd think that the beta=0.5 is the better choice for this dataset. Reasons:

- markedly higher rewards margins compared to all other betas
- better balance between positive chosen and negative rejected rewards
- log probabilities are not as superbly low as for beta=0.01, which seems too low for this dataset

Of course, that is just purely looking at numbers without running any benchmarks. However, I am hesitant to evaluate all the models on benchmarks and, therefore, literally optimising my hyperparameters on a test set (which is very bad!). So I will just play with some of the most promising models and see which one feels "best" qualitatively.

If you have other insights, thoughts, or opinions, let me know!
  • 3 replies
·
BramVanroy 
posted an update 11 months ago
view post
Post
🕵️ Looking for DPO experts!

I have a dataset that is gpt-4-turbo as chosen and a lower performing model as rejected. The objective should therefore be fairly easy because the two are easy to discern. As a consequence, the model achieves very low losses (0.021 train; 0.013 validation) and high reward accuracies (0.995). **However**, when using the model in practice, it often detoriates after the first one or two tokens and continuously outputs sequences of /*****/. So despite the good performance on the DPO objective and strong scores on the validation set (no overfitting), something seems to go wrong. Perhaps the outputs are too different and the task is too easy, in which case DPO is not useful. But why then would the model start hallucinating and repeating the same token over and over again?

Any thoughts? Any suggestions to get around this? All discussions are welcome!
·
BramVanroy 
posted an update 11 months ago
view post
Post
💡We recently launched a Discord server for #Dutch #NLP and #LLMs. We have more than 50 users of _very_ varrying backgrounds! 🧙👩‍🔬🧑‍🎨🧑‍🏫🧑‍💼 We've already had discussions on eval, tokenizers, RAG, data... A bit of everything. Everyone is welcome to work together on Dutch NLP and LLMs! https://discord.gg/YUUXVZkZJ9
·
BramVanroy 
authored a paper about 1 year ago