Multiple llama4 fixe (#37353)

* update for fixes * more fixes * fuxix dynamic cache? * style * fix both traiining and generating. Eager seems alright * dynamic does not work * fix most cases, use_cache or not, eager or not, no default cache (ex: not training but you want to get cache states) * should be final fixes * fix more stuff no cat * style * fix * style * final sytle * qualityeioiwhjfaopsejdpofqsdjkfjha;wesdhgfkjlqsw.denghjkaswednkgs * fix * revert
2025-04-08 11:14:49 +02:00
parent 794fde7b1c
commit 2da82e432d
7 changed files with 65 additions and 48 deletions
--- a/utils/check_config_attributes.py
+++ b/utils/check_config_attributes.py
@@ -244,6 +244,7 @@ SPECIAL_CASES_TO_ALLOW = {
        "output_router_logits",
        "router_aux_loss_coef",
        "router_jitter_noise",
+        "cache_implementation",
    ],
    "Llama4VisionConfig": ["multi_modal_projector_bias", "norm_eps"],
 }