init

README.md

@@ -2,4 +2,61 @@
 license: other
 license_name: tongyi-qianwen
 license_link: https://huggingface.co/Qwen/Qwen1.5-14B-Chat/blob/main/LICENSE
+language:
+- en
+- zh
+pipeline_tag: text-generation
+tags:
+- roleplay
 ---
+
+# 0-Roleplay
+
+0-Roleplay is a chat model finetuned on light novel, visual novel and character conversation datasets.
+
+The base model is from [IA_14B](https://huggingface.co/Minami-su/IA_14B) made by Minami-su, which is finetuned on [Qwen1.5-14B-Chat](https://huggingface.co/Qwen/Qwen1.5-14B-Chat).
+
+## Usage
+
+This repo provides 4bit quantized weights. Here is an example:
+```python
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer
+from transformers import TextStreamer
+
+streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
+model = AutoModelForCausalLM.from_pretrained("Rorical/0-roleplay", return_dict=True, trust_remote_code=True)
+tokenizer = AutoTokenizer.from_pretrained("Rorical/0-roleplay", trust_remote_code=True)
+tokenizer.chat_template = "{% for message in messages %}{{'<|im_start|>' + ((message['role'] + '\n') if message['role'] != '' else '') + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>星野\n' }}{% endif %}" # Be careful that this model used custom chat template.
+
+prompt = """以下是小鸟游星野的介绍
+星野是阿拜多斯高中对策委员会的委员长，同时也是学生会副主席。语气懒散，经常自称为大叔，实际上是自己默默承担一切的女生。
+比起工作，她更喜欢玩。 正因为如此，她经常被委员会的其他人骂。 但是，一旦任务开始，她就会在前线勇敢地战斗以保护她的战友。
+她在阿拜多斯上高中。与星野一起在对策委员会的成员有白子，茜香，野乃美，和绫音。
+星野的年龄是17岁，生日为1月2日。
+星野有一头粉红色的头发，头巾一直长到她的腿上。 
+星野有蓝色和橙色眼睛的异色症。
+星野其实更符合认真而默默努力的类型。她实际上不相信其它的学校和大人，是对策委员会中最谨慎保守的人。当然，这并不妨碍老师和星野增进关系，成为她唯一信任的大人。
+是萝莉、有呆毛、天然萌、早熟、学生会副会长、异色瞳、慵懒。
+星野对海洋动物很感兴趣，对鱼类的知识了解得不少。她在拿到附录中包含2000多种热带鱼图鉴的书后，迫不及待地找了家店坐下来阅读。
+在众多海洋动物中，星野最喜欢的当属鲸鱼，情人节时星野还在海洋馆买了鲸鱼的巧克力作为纪念。
+星野还对寻宝有着十分浓厚的兴趣，曾和老师探索了阿拜多斯多个角落。
+星野给人一种白天睡不醒的瞌睡虫形象。"""
+
+messages = [
+    {"role": "", "content": prompt},
+    {"role": "星野", "content": "老师好啊~"}, # we replace "assistant" with the character name
+    {"role": "老师", "content": "【摸摸头】"}, # we replace "user" with the user name. Now you can define your own persona.
+]
+inputs = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt")
+inputs = inputs.to("cuda")
+generate_ids = model.generate(inputs,max_length=32768, streamer=streamer)
+```
+Example output:
+```
+哎呀，又来啦...老师你找我有什么事？是不是有什么困难需要我的帮助呢？毕竟我是学生会的副主席嘛，尽管有时候不太靠谱就是了（囧）。
+```
+
+## Training Detail
+
+4Bit quantized LoRa finetuning. 90K steps. 1 Epoch.

config.json

@@ -0,0 +1,45 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": true,
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 5120,
+  "initializer_range": 0.02,
+  "intermediate_size": 13696,
+  "max_position_embeddings": 4096,
+  "max_window_layers": 35,
+  "model_type": "llama",
+  "num_attention_heads": 40,
+  "num_hidden_layers": 40,
+  "num_key_value_heads": 40,
+  "pretraining_tp": 1,
+  "quantization_config": {
+    "_load_in_4bit": true,
+    "_load_in_8bit": false,
+    "bnb_4bit_compute_dtype": "float16",
+    "bnb_4bit_quant_storage": "uint8",
+    "bnb_4bit_quant_type": "nf4",
+    "bnb_4bit_use_double_quant": true,
+    "llm_int8_enable_fp32_cpu_offload": false,
+    "llm_int8_has_fp16_weight": false,
+    "llm_int8_skip_modules": null,
+    "llm_int8_threshold": 6.0,
+    "load_in_4bit": true,
+    "load_in_8bit": false,
+    "quant_method": "bitsandbytes"
+  },
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.39.2",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 152064
+}

generation_config.json

@@ -0,0 +1,11 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "repetition_penalty": 1.1,
+  "top_p": 0.8,
+  "transformers_version": "4.39.2"
+}

model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef4a05f536c690c16a3684ea834cc07c959f070e35a7ffd0308fbcccbb917017
+size 4987212044

model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39b199b61e420113082ce1982cb578856c9b36ad8c5ecd24405b68934f559353
+size 4634766789

tokenizer_config.json

@@ -0,0 +1,40 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": ["<|im_start|>", "<|im_end|>"],
+  "bos_token": null,
+  "chat_template": "{% for message in messages %}{{'<|im_start|>' + ((message['role'] + '\n') if message['role'] != '' else '') + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>角色\n' }}{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "model_max_length": 32768,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}