@@ -175,6 +175,7 @@ class llama_token_data_array(Structure):
175175# // context pointer passed to the progress callback
176176# void * progress_callback_user_data;
177177
178+
178179# // Keep the booleans together to avoid misalignment during copy-by-value.
179180# bool low_vram; // if true, reduce VRAM usage at the cost of performance
180181# bool f16_kv; // use fp16 for KV cache
@@ -748,7 +749,12 @@ def llama_get_vocab(
748749 return _lib .llama_get_vocab (ctx , strings , scores , capacity )
749750
750751
751- _lib .llama_get_vocab .argtypes = [llama_context_p , Array [c_char_p ], Array [c_float ], c_int ]
752+ _lib .llama_get_vocab .argtypes = [
753+ llama_context_p ,
754+ POINTER (c_char_p ),
755+ POINTER (c_float ),
756+ c_int ,
757+ ]
752758_lib .llama_get_vocab .restype = c_int
753759
754760
@@ -766,6 +772,15 @@ def llama_get_vocab_from_model(
766772 return _lib .llama_get_vocab_from_model (model , strings , scores , capacity )
767773
768774
775+ _lib .llama_get_vocab_from_model .argtypes = [
776+ llama_model_p ,
777+ POINTER (c_char_p ),
778+ POINTER (c_float ),
779+ c_int ,
780+ ]
781+ _lib .llama_get_vocab_from_model .restype = c_int
782+
783+
769784# Token logits obtained from the last call to llama_eval()
770785# The logits for the last token are stored in the last row
771786# Can be mutated in order to change the probabilities of the next token
0 commit comments