Skip to content

GitLab

Switch branch/tag
  1. 16 Mar, 2021 1 commit
  3. 15 Mar, 2021 16 commits
  5. 12 Mar, 2021 7 commits
    • Stas Bekman's avatar
      AdamW is now supported by default (#9624) · 4c32f9f2
      Stas Bekman authored
      4c32f9f2
    • ymfa's avatar
      Pass encoder outputs into GenerationMixin (#10599) · fa35cda9
      ymfa authored 
      * Pass encoder_outputs into generate()

* Remove an if-statement

* Reformat

* Minimize changes to generate()

* Comment on input_ids
      fa35cda9
    • PaulLerner's avatar
      fix: #10628 expanduser path in TrainingArguments (#10660) · 00cad2e5
      PaulLerner authored 
      

* fix: #10628 expanduser path in TrainingArguments

* docs: explain why we expand paths in TrainingArguments

* Style
Co-authored-by: default avatarSylvain Gugger <sylvain.gugger@gmail.com>
      00cad2e5
    • Sylvain Gugger's avatar
      Add auto_wrap option in fairscale integration (#10673) · e8246f78
      Sylvain Gugger authored 
      * Add auto_wrap option in fairscale integration

* Style
      e8246f78
    • Lysandre Debut's avatar
      TensorFlow tests: having from_pt set to True requires torch to be installed. (#10664) · 184ef8ec
      Lysandre Debut authored 
      * TF model exists for Blenderbot 400M

* Marian

* RAG
      184ef8ec
    • Nicolas Patry's avatar
      Adding new parameter to `generate`: `max_time`. (#9846) · 543d0549
      Nicolas Patry authored 
      * [WIP] Adding new parameter to `generate`:  `max_time`.

Generation by tokens number is sometimes a bit clunky because we don't
know how many tokens are good enough or even how many tokens are in
the payload (for pipelines users for instance). This leads to hard
to understand behavior.

This PR proposes a new argument `max_time` which is a float of seconds
for the allowed time for `generate` to run on.
Ideally combinations of `max_tokens=None`, `max_time=2` could be used to
generate as many tokens as possible within time budget.

NB: Another possible approach consists of passing a callback to `generate`
  putting the caller in charge of the actual decision of when to stop
  generating tokens. It opens the door to 'which args should we pass'
  to this callback. It's hard to imagine other use-cases for this
  early stopping behavior than time (that are not already covered by
  parameters of generate)

* Revamp with StoppingCriteria

* Removing deprecated mentions.

* Forgot arguments to stopping criteria.

* Readding max_length it's not just used as a stopping criteria.

* Default value for `stopping_criteria`.

* Address @patrickvonplaten comments.

- More docstrings
- Actual doc
- Include in global namespace
- Remove TF work.

* Put back `max_length` (deprecation different PR).

* Doc quality.

* Fixing old behavior without `stopping_criteria` but with `max_length`.

Making sure we don't break that in the future.

* Adding more tests for possible inconsistencies between

`max_length` and `stopping_criteria`.

* Fixing the torch imports.
      543d0549
    • Lysandre Debut's avatar
      Adjust loss difference (#10669) · ea46e3fa
      Lysandre Debut authored
      ea46e3fa
  7. 11 Mar, 2021 16 commits