Comments (6)
Residual value prediction didn't help with stability (it's crimson-wish)
from trl.
Also not a big help via the other approx KL formulation. W&B here. Though, it's slightly more stable?
We'll see how this run finishes converging.
from trl.
It could be good to make things like this configurable in a branch and learning how these implementation details transfer to RLHF.
from trl.
imo, residual clipping seems beneficial to prevent policy loss spiking reported in #101 . It's probably coming from instability in value estimation.
from trl.
Yeah, I'm running residual clipping example(s), we'll see. At least it'll be good to have the option to try both.
from trl.
Closing this for now, feel free to reopen if there's an update.
from trl.
Related Issues (20)
- KTO train_loss = 0.0 HOT 11
- FSDP Must flatten tensors with uniform dtype but got torch.bfloat16 and torch.float32 HOT 3
- FSDP with PPO trainer won't work because FSDP doesn't support model.generate HOT 7
- Questions about the reference model in PPOTrainer and DPOTrainer HOT 5
- [BUG] RuntimeError: still have inflight params in KTO
- TRLParser needs changes, overwrites command line arguments with config
- TRL CLI leads to issue in DPO when config and cli arguments are provided HOT 4
- DPOTrainer: AttributeError: 'list' object has no attribute 'numel' HOT 2
- TypeError when not passing total_episodes in PPOv2Trainer HOT 4
- TypeError: IterableDataset.map() got an unexpected keyword argument 'num_proc' with streaming datasets HOT 1
- [DOCS] Some headers (h2, h3) not rendered properly
- DPODataCollatorWithPadding doesn't mask the prompts
- [BUG] TRL CLI not capturing `torch_dtype` correctly HOT 1
- Question about apply_chat_template in examples HOT 7
- Misuse of gen_len in examples/notebooks/gpt2-sentiment.ipynb
- DataCollatorForCompletionOnlyLM does not work with FSDP
- PPOv2 trainer, the wandb log is unnormal HOT 8
- What is the difference between PPOv2Trainer and PPOTrainer? HOT 1
- Using IterableDataset crashed the SFTTrainer
- Neftune is applied twice; in trl and transformers BOTH! HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from trl.