Add tokenizer and update model

README.md

@@ -0,0 +1,28 @@
+# Korean GPT
+
+한국어 GPT 모델입니다.
+
+## 사용법
+```python
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+tokenizer = AutoTokenizer.from_pretrained(
+    "oz1115/korean-gpt-quick-test",
+    trust_remote_code=True
+)
+
+model = AutoModelForCausalLM.from_pretrained(
+    "oz1115/korean-gpt-quick-test",
+    trust_remote_code=True
+)
+
+inputs = tokenizer("안녕하세요", return_tensors="pt")
+outputs = model.generate(**inputs, max_length=50)
+print(tokenizer.decode(outputs[0]))
+```
+
+## 모델 정보
+- Vocabulary: 32,000
+- Hidden Size: 512
+- Layers: 8
+- Attention Heads: 8

config.json

@@ -0,0 +1,29 @@
+{
+  "architectures": [
+    "KoreanGPTModel"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 2,
+  "dtype": "float32",
+  "eos_token_id": 3,
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 256,
+  "initializer_range": 0.02,
+  "intermediate_size": 1024,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 256,
+  "model_type": "korean_gpt",
+  "num_attention_heads": 4,
+  "num_hidden_layers": 4,
+  "pad_token_id": 0,
+  "transformers_version": "4.57.3",
+  "use_cache": true,
+  "vocab_size": 32000,
+  "auto_map": {
+    "AutoModelForCausalLM": "modeling_korean_gpt.KoreanGPTModel",
+    "AutoTokenizer": [
+      "tokenization_korean_gpt.KoreanGPTTokenizer",
+      null
+    ]
+  }
+}

korean_sp_32k.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3ed6db693f545bd90ef9fed4d6415a511a7f59ac168c7f5e757a75c06d704fa
+size 836193

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd8b533a5c3e7b5af23c7587f0d2085543f8e0c0634610e7dc8dd2a3447ac7e0
+size 46739615

special_tokens_map.json

@@ -0,0 +1,6 @@
+{
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "unk_token": "<unk>",
+  "pad_token": "<pad>"
+}

tokenizer_config.json

@@ -0,0 +1,14 @@
+{
+  "tokenizer_class": "KoreanGPTTokenizer",
+  "auto_map": {
+    "AutoTokenizer": [
+      "tokenization_korean_gpt.KoreanGPTTokenizer",
+      null
+    ]
+  },
+  "model_max_length": 512,
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "unk_token": "<unk>",
+  "pad_token": "<pad>"
+}