- 14 Jul, 2024 1 commit
-
-
Joao Gante authored
* tmp commit * shorter * nit * explicit kwargs * propagate changes * mass propagation with a few manual touches (let's see how CI behaves) * fix cacheless case * Update src/transformers/generation/utils.py Co-authored-by:
Arthur <48595927+ArthurZucker@users.noreply.github.com> * make fixup --------- Co-authored-by:
-
- 02 Jul, 2024 1 commit
-
-
Yih-Dar authored
* move * move * move * move * Update tests/utils/test_image_processing_utils.py Co-authored-by:
amyeroberts <22614925+amyeroberts@users.noreply.github.com> --------- Co-authored-by:
ydshieh <ydshieh@users.noreply.github.com> Co-authored-by:
-
- 26 Jun, 2024 1 commit
-
-
amyeroberts authored
* Skip tests properly * [test_all] * Add 'reason' as kwarg for skipTest * [test_all] Fix up * [test_all]
-
- 06 Mar, 2024 1 commit
-
-
Joao Gante authored
-
- 20 Feb, 2024 1 commit
-
-
Joao Gante authored
-
- 15 Feb, 2024 1 commit
-
-
Arthur authored
* wow I was scared! * fix everything * nits * make it BC? * add todo * nits * is_tracing should still be used to pass tracing tests * nits * some nits to make sure genration works with static cache uncompiled * fix sdpa * fix FA2 for both static and dynamic in a better way? * style * fix-copies * fix fix copies * fix sequential beam searcg * style * use `keys_to_ignore` * nit * correct dtype inference when init * :( the fix for FA2 is still not optimal to investigate! * styling * nits * nit * this might work better * add comment * Update src/transformers/models/llama/modeling_llama.py * "position_ids" -> "cache_position" * style * nit * Remove changes that should no be propagatted just yet * Apply suggestions from code review * Styling * make sure we raise an errir for static cache with FA2 enabled * move to the bottom of the signature * style * Update src/transformers/models/llama/modeling_llama.py Co-authored-by:
Younes Belkada <49240599+younesbelkada@users.noreply.github.com> * Update src/transformers/models/llama/modeling_llama.py * nit in the name --------- Co-authored-by:
-
- 13 Feb, 2024 1 commit
-
-
Joao Gante authored
-
- 08 Feb, 2024 1 commit
-
-
Arthur authored
Co-authored-by:
fxmarty <9808326+fxmarty@users.noreply.github.com> Co-authored-by:
Joao Gante <joaofranciscocardosogante@gmail.com>
-
- 08 Dec, 2023 2 commits
-
-
Joao Gante authored
-
Tom Aarsen authored
* Draft version of new KV Caching This should allow Attention Sinks (https://github.com/tomaarsen/attention_sinks) / StreamingLLM (https://arxiv.org/abs/2309.17453) to be easily implemented in a third-party or in transformers directly * Address numerous PR suggestions 1. Move layer_idx from cache to ...Attention. Removes confusing set_layer_idx magic. 2. Always convert past_key_values to Cache instance at the start of ...Attention, removes all other isinstance calls. 3. Remove __bool__ and __getitem__ magic as they're confusing. 4. past_key_values.update(key, value, idx) now returns key, value. 5. Add use_legacy_cache flag, defaults to None, i.e. Falsey. This breaks generate for now, until 1) the cache is used is generate() or 2) use_legacy_cache is defaulted to True in generate() until we change it in another PR. 6. Separate key_cache and value_cache. Some work is still needed to see if the SinkCache can conveniently be implemented with just one update method. * Implement the SinkCache through backward+forward rotations * Integrate (Sink)Cache with Llama FA2 * Set use_legacy_cache=True as default, allows for test passes * Move from/to_legacy_cache to ...Model class * Undo unnecessary newline change * Remove copy utility from deprecated OpenLlama * Match import style * manual rebase with main * Cache class working with generate (#1) * Draft version of new KV Caching This should allow Attention Sinks (https://github.com/tomaarsen/attention_sinks) / StreamingLLM (https://arxiv.org/abs/2309.17453 ) to be easily implemented in a third-party or in transformers directly * Address numerous PR suggestions 1. Move layer_idx from cache to ...Attention. Removes confusing set_layer_idx magic. 2. Always convert past_key_values to Cache instance at the start of ...Attention, removes all other isinstance calls. 3. Remove __bool__ and __getitem__ magic as they're confusing. 4. past_key_values.update(key, value, idx) now returns key, value. 5. Add use_legacy_cache flag, defaults to None, i.e. Falsey. This breaks generate for now, until 1) the cache is used is generate() or 2) use_legacy_cache is defaulted to True in generate() until we change it in another PR. 6. Separate key_cache and value_cache. Some work is still needed to see if the SinkCache can conveniently be implemented with just one update method. * Integrate (Sink)Cache with Llama FA2 * Move from/to_legacy_cache to ...Model class * Undo unnecessary newline change * Match import style * working generate * Add tests; Simplify code; Apply changes to Mistral and Persimmon * fix rebase mess * a few more manual fixes * last manual fix * propagate changes to phi * upgrade test * add use_legacy_cache docstring; beef up tests * reintroduce unwanted deletes --------- Co-authored-by:
Tom Aarsen <Cubiegamedev@gmail.com> * move import * add default to model_kwargs.get('use_legacy_cache') * correct failing test * Apply suggestions from code review Co-authored-by:
Patrick von Platen <patrick.v.platen@gmail.com> * apply PR suggestions * fix failing test * Apply suggestions from code review Co-authored-by:
Tom Aarsen <37621491+tomaarsen@users.noreply.github.com> * PR comments * tmp commit * add docstrings * more tests, more docstrings, add to docs * derp * tmp commit * tmp dbg * more dbg * fix beam search bug * cache can be a list of tuples in some models * fix group beam search * all but sinkcache integration tests * fix sink cache and add hard integration test * now also compatible with input_embeds input * PR comments * add Cache support to Phi+FA2 * make fixup --------- Co-authored-by:
Joao Gante <joao@huggingface.co> Co-authored-by:
-
