DeepSpeed
88ba24a3 - Use past_key_value when provided (#7428)

Commit
166 days ago
Use past_key_value when provided (#7428) The KV cache can be passed via `layer_past` or `past_key_value` arguments. Previously, `past_key_value` was ignored, causing workload incompatibilities. This PR fixes the issue while preserving the original logic. --------- Signed-off-by: Max Kovalenko <mkovalenko@habana.ai>
Author
Parents
Loading