Load GLM vision weights on demand

This commit is contained in:
Georg Bauer
2026-09-01 20:08:36 +02:00
parent 46d6a976a5
commit ee515ee824
10 changed files with 293 additions and 26 deletions

View File

@@ -118,6 +118,10 @@ int ds4_gpu_build_derived_artifacts(const void *model_map, uint64_t model_size,
int ds4_gpu_model_range_replaced(const void *model_map, uint64_t offset,
uint64_t bytes);
int ds4_gpu_set_model_map_range(const void *model_map, uint64_t model_size, uint64_t map_offset, uint64_t map_size, uint64_t max_tensor_bytes);
int ds4_gpu_set_transient_model_map_range(const void *model_map, uint64_t model_size, uint64_t map_offset, uint64_t map_size, uint64_t max_tensor_bytes);
/* Caller must finish every command that references this mapping first. */
int ds4_gpu_release_transient_model_map(const void *model_map, uint64_t model_size);
int ds4_gpu_model_map_active(const void *model_map, uint64_t model_size);
/* Add a secondary GGUF mapping without replacing the primary model mapping. */
int ds4_gpu_set_aux_model_map_range(const void *model_map,
uint64_t model_size,