Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
ejddhbrbrrnrn
84 days ago
|
parent
|
context
|
favorite
| on:
Bugs in LLM Training – Gradient Accumulation Fix
Is this a general issue rather than unsloth specific. How wide is this problem? Sounds wild if it has been affecting everyones training.
danielhanchen
84 days ago
[–]
Unfortunately it's not an Unsloth issue but a general issue affecting nearly all trainers which use grad accum. We worked with Huggingface so their trainers should be fixed now though in the main branch
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: