More model refactoring! (#35359)

* cohere

* style

* phi3

* style

* small fix

* small fix

* phi3 longrope

* oups

* Update rope (only for phi3 still)

* Update test_modeling_rope_utils.py

* Update modeling_phi3.py

* fix

* fix copies

* style

* Fix copied from bad renaming
This commit is contained in:
Cyril Vallez
2025-01-09 11:09:09 +01:00
committed by GitHub
parent 137965ca7d
commit 965a2fb320
36 changed files with 1253 additions and 1243 deletions

View File

@@ -51,6 +51,8 @@ SPECIAL_CASES_TO_ALLOW = {
# generation configs (TODO joao)
"Gemma2Config": ["tie_word_embeddings", "cache_implementation"],
"Cohere2Config": ["cache_implementation"],
# Dropout with this value was declared but never used
"Phi3Config": ["embd_pdrop"],
# used to compute the property `self.chunk_length`
"EncodecConfig": ["overlap"],
# used to compute the property `self.layers_block_type`