Create working RadonSAI with real weights

README.md

@@ -3,13 +3,19 @@ license: apache-2.0
 tags:
 - radon
 - gpt2
-- 2000mb
-- fixed
+- working
+- 1024d
 ---
 
-# RadonSAI (Fixed)
+# RadonSAI (Working)
 
-Исправленная версия RadonSAI с рабочей конфигурацией.
+Полностью рабочая версия RadonSAI с реальными весами.
+
+## Технические характеристики
+- Параметры: ~203,668,480
+- Размерность: 1024
+- Слои: 12
+- Головы внимания: 16
 
 ## Использование
 ```python
@@ -17,4 +23,26 @@ from transformers import AutoModelForCausalLM, AutoTokenizer
 
 model = AutoModelForCausalLM.from_pretrained("MagistrTheOne/RadonSAI")
 tokenizer = AutoTokenizer.from_pretrained("MagistrTheOne/RadonSAI")
+
+# Генерация текста
+input_text = "Hello, how are you?"
+inputs = tokenizer(input_text, return_tensors="pt")
+outputs = model.generate(**inputs, max_length=50)
+result = tokenizer.decode(outputs[0], skip_special_tokens=True)
+print(result)
+```
+
+## Тестирование
+```python
+# Быстрый тест
+import torch
+model = AutoModelForCausalLM.from_pretrained("MagistrTheOne/RadonSAI")
+tokenizer = AutoTokenizer.from_pretrained("MagistrTheOne/RadonSAI")
+
+# Простая генерация
+text = "The future of AI is"
+inputs = tokenizer(text, return_tensors="pt")
+with torch.no_grad():
+    outputs = model.generate(**inputs, max_length=30, do_sample=True, temperature=0.7)
+print(tokenizer.decode(outputs[0], skip_special_tokens=True))
 ```

config.json

@@ -1,28 +1,31 @@
 {
+  "activation_function": "gelu_new",
   "architectures": [
     "GPT2LMHeadModel"
   ],
-  "model_type": "gpt2",
-  "n_ctx": 1024,
-  "n_embd": 1024,
-  "n_head": 16,
-  "n_layer": 12,
-  "n_positions": 1024,
-  "vocab_size": 50257,
-  "torch_dtype": "float16",
-  "transformers_version": "4.36.2",
-  "use_cache": true,
-  "attention_dropout": 0.0,
   "attn_pdrop": 0.1,
   "bos_token_id": 50256,
-  "eos_token_id": 50256,
   "embd_pdrop": 0.1,
+  "eos_token_id": 50256,
   "initializer_range": 0.02,
   "layer_norm_epsilon": 1e-05,
+  "model_type": "gpt2",
+  "n_embd": 1024,
+  "n_head": 16,
+  "n_inner": null,
+  "n_layer": 12,
+  "n_positions": 1024,
+  "reorder_and_upcast_attn": false,
   "resid_pdrop": 0.1,
+  "scale_attn_by_inverse_layer_idx": false,
+  "scale_attn_weights": true,
   "summary_activation": null,
   "summary_first_dropout": 0.1,
   "summary_proj_to_labels": true,
   "summary_type": "cls_index",
-  "summary_use_proj": true
-}
+  "summary_use_proj": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.36.2",
+  "use_cache": true,
+  "vocab_size": 50257
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 50256,
+  "eos_token_id": 50256,
+  "transformers_version": "4.36.2"
+}

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3323be5474d086a52bf5c73dee21b5438b501c1f4b007342edbe6cada51e25c9
-size 131278312
+oid sha256:9e4cb54b9ebcdc8a72ebb3265c871023c38189e7d6851a87641fe6aeafce66ca
+size 814689096

special_tokens_map.json

@@ -13,6 +13,7 @@
     "rstrip": false,
     "single_word": false
   },
+  "pad_token": "<|endoftext|>",
   "unk_token": {
     "content": "<|endoftext|>",
     "lstrip": false,

tokenizer_config.json

@@ -1,14 +1,22 @@
 {
-  "auto_map": {
-    "AutoTokenizer": [
-      "gpt2",
-      null
-    ]
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "50256": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
   },
   "bos_token": "<|endoftext|>",
+  "clean_up_tokenization_spaces": true,
   "eos_token": "<|endoftext|>",
+  "errors": "replace",
   "model_max_length": 1024,
   "pad_token": "<|endoftext|>",
   "tokenizer_class": "GPT2Tokenizer",
   "unk_token": "<|endoftext|>"
-}
+}