Update README.md

README.md

@@ -16,14 +16,14 @@ tags:
 
 InfCapital LLama2-7b clone of [Meta's Llama 2 7B Chat](https://huggingface.co/meta-llama/Llama-2-7b-chat-hf).
 
-Adopt for Vietnamese continued pretraining or fine-tuning by extend vocab size from 32,000 to 48,281
+Adopt for Vietnamese continued pretraining or fine-tuning by extend vocab size from 32,000 to 44,800. Vocabs added by training sentencepiece method from dataset vnnews-corpus.
 
 
 ## Model Architecture
 ```
 LlamaForCausalLM(
   (model): LlamaModel(
-    (embed_tokens): Embedding(32000, 4096, padding_idx=0)
+    (embed_tokens): Embedding(44800, 4096)
     (layers): ModuleList(
       (0-31): 32 x LlamaDecoderLayer(
         (self_attn): LlamaAttention(
@@ -45,6 +45,6 @@ LlamaForCausalLM(
     )
     (norm): LlamaRMSNorm()
   )
-  (lm_head): Linear(in_features=4096, out_features=32000, bias=False)
+  (lm_head): Linear(in_features=4096, out_features=44800, bias=False)
 )
 ```