convert : refactor vocab selection logic (ggml-org#6355)

Author: cebtenzzre

convert-hf-to-gguf.py

@@ -23,7 +23,7 @@
     sys.path.insert(1, str(Path(__file__).parent / 'gguf-py'))
 import gguf
 
-from convert import HfVocab
+from convert import LlamaHfVocab
 
 
 ###### MODEL DEFINITIONS ######
@@ -230,7 +230,7 @@ def _get_part_names(self):
     def _set_vocab_gpt2(self):
         dir_model = self.dir_model
         hparams = self.hparams
-        tokens: list[bytearray] = []
+        tokens: list[str] = []
         toktypes: list[int] = []
 
         from transformers import AutoTokenizer
@@ -243,8 +243,7 @@ def _set_vocab_gpt2(self):
 
         for i in range(vocab_size):
             if i not in reverse_vocab:
-                pad_token = f"[PAD{i}]".encode('utf-8')
-                tokens.append(bytearray(pad_token))
+                tokens.append(f"[PAD{i}]")
                 toktypes.append(gguf.TokenType.USER_DEFINED)
             elif reverse_vocab[i] in added_vocab:
                 tokens.append(reverse_vocab[i])
@@ -266,7 +265,7 @@ def _set_vocab_gpt2(self):
     def _set_vocab_qwen(self):
         dir_model = self.dir_model
         hparams = self.hparams
-        tokens: list[bytearray] = []
+        tokens: list[str] = []
         toktypes: list[int] = []
 
         from transformers import AutoTokenizer
@@ -291,8 +290,7 @@ def _set_vocab_qwen(self):
 
         for i in range(vocab_size):
             if i not in reverse_vocab:
-                pad_token = f"[PAD{i}]".encode("utf-8")
-                tokens.append(bytearray(pad_token))
+                tokens.append(f"[PAD{i}]")
                 toktypes.append(gguf.TokenType.USER_DEFINED)
             elif reverse_vocab[i] in added_vocab:
                 tokens.append(reverse_vocab[i])
@@ -372,12 +370,8 @@ def _set_vocab_sentencepiece(self):
         special_vocab = gguf.SpecialVocab(self.dir_model, n_vocab=len(tokens))
         special_vocab.add_to_gguf(self.gguf_writer)
 
-    def _set_vocab_hf(self):
-        path = self.dir_model
-        added_tokens_path = self.dir_model
-        vocab = HfVocab(
-            path, added_tokens_path if added_tokens_path.exists() else None
-        )
+    def _set_vocab_llama_hf(self):
+        vocab = LlamaHfVocab(self.dir_model)
         tokens = []
         scores = []
         toktypes = []
@@ -1099,7 +1093,7 @@ def set_gguf_parameters(self):
         self.gguf_writer.add_file_type(self.ftype)
 
     def set_vocab(self):
-        self._set_vocab_hf()
+        self._set_vocab_llama_hf()
 
     def _reverse_hf_permute(self, weights: Tensor, n_head: int, n_kv_head: int | None = None) -> Tensor:
         if n_kv_head is not None and n_head != n_kv_head:
@@ -1700,11 +1694,8 @@ def set_gguf_parameters(self):
             self.gguf_writer.add_pooling_type(pooling_type)
 
     def set_vocab(self):
-        path = self.dir_model
-        added_tokens_path = self.dir_model if self.dir_model.exists() else None
-
         # use huggingface vocab to get all tokens
-        vocab = HfVocab(path, added_tokens_path)
+        vocab = LlamaHfVocab(self.dir_model, ignore_nonllama=True)
         tokens, scores, toktypes = zip(*vocab.all_tokens())
         assert len(tokens) == vocab.vocab_size
         self.vocab_size = vocab.vocab_size

convert-persimmon-to-gguf.py

@@ -106,12 +106,12 @@ def main():
     tensor_map = gguf.get_tensor_name_map(arch, block_count)
     print(tensor_map)
     for name in tensors.keys():
-        data = tensors[name]
+        data_torch = tensors[name]
         if name.endswith(".self_attention.rotary_emb.inv_freq"):
             continue
-        old_dtype = data.dtype
+        old_dtype = data_torch.dtype
         # TODO: FP16 conversion produces garbage outputs. (Q8_0 does not, so..?)
-        data = data.to(torch.float32).squeeze().numpy()
+        data = data_torch.to(torch.float32).squeeze().numpy()
         new_name = tensor_map.get_name(name, try_suffixes = (".weight", ".bias"))
         if new_name is None:
             print("Can not map tensor '" + name + "'")