first commit

README.md

@@ -1,3 +1,173 @@
-# rwkv-5-world-1b5_a13650644885172224812389
+### Run Huggingface RWKV5 World Model
 
-rwkv-5-world-1b5
+
+#### CPU
+
+```python
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+def generate_prompt(instruction, input=""):
+    instruction = instruction.strip().replace('\r\n','\n').replace('\n\n','\n')
+    input = input.strip().replace('\r\n','\n').replace('\n\n','\n')
+    if input:
+        return f"""Instruction: {instruction}
+
+Input: {input}
+
+Response:"""
+    else:
+        return f"""User: hi
+
+Assistant: Hi. I am your assistant and I will provide expert full response in full details. Please feel free to ask any question and I will always answer it.
+
+User: {instruction}
+
+Assistant:"""
+
+
+model = AutoModelForCausalLM.from_pretrained("RWKV/rwkv-5-world-1b5", trust_remote_code=True).to(torch.float32)
+tokenizer = AutoTokenizer.from_pretrained("RWKV/rwkv-5-world-1b5", trust_remote_code=True, padding_side='left', pad_token="<s>")
+
+text = "请介绍北京的旅游景点"
+prompt = generate_prompt(text)
+
+inputs = tokenizer(prompt, return_tensors="pt")
+output = model.generate(inputs["input_ids"], max_new_tokens=333, do_sample=True, temperature=1.0, top_p=0.3, top_k=0, )
+print(tokenizer.decode(output[0].tolist(), skip_special_tokens=True))
+```
+
+output:
+
+```shell
+User: hi
+
+Assistant: Hi. I am your assistant and I will provide expert full response in full details. Please feel free to ask any question and I will always answer it.
+
+User: 请介绍北京的旅游景点
+
+Assistant: 北京是中国的首都，拥有众多的旅游景点，以下是其中一些著名的景点：
+1. 故宫：位于北京市中心，是明清两代的皇宫，内有大量的文物和艺术品。
+2. 天安门广场：是中国最著名的广场之一，是中国人民政治协商会议的旧址，也是中国人民政治协商会议的中心。
+3. 颐和园：是中国古代皇家园林之一，有着悠久的历史和丰富的文化内涵。
+4. 长城：是中国古代的一道长城，全长约万里，是中国最著名的旅游景点之一。
+5. 北京大学：是中国著名的高等教育机构之一，有着悠久的历史和丰富的文化内涵。
+6. 北京动物园：是中国最大的动物园之一，有着丰富的动物资源和丰富的文化内涵。
+7. 故宫博物院：是中国最著名的博物馆之一，收藏了大量的文物和艺术品，是中国最重要的文化遗产之一。
+8. 天坛：是中国古代皇家
+```
+
+#### GPU
+
+```python
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+def generate_prompt(instruction, input=""):
+    instruction = instruction.strip().replace('\r\n','\n').replace('\n\n','\n')
+    input = input.strip().replace('\r\n','\n').replace('\n\n','\n')
+    if input:
+        return f"""Instruction: {instruction}
+
+Input: {input}
+
+Response:"""
+    else:
+        return f"""User: hi
+
+Assistant: Hi. I am your assistant and I will provide expert full response in full details. Please feel free to ask any question and I will always answer it.
+
+User: {instruction}
+
+Assistant:"""
+
+
+model = AutoModelForCausalLM.from_pretrained("RWKV/rwkv-5-world-1b5", trust_remote_code=True, torch_dtype=torch.float16).to(0)
+tokenizer = AutoTokenizer.from_pretrained("RWKV/rwkv-5-world-1b5", trust_remote_code=True, padding_side='left', pad_token="<s>")
+
+text = "介绍一下大熊猫"
+prompt = generate_prompt(text)
+
+inputs = tokenizer(prompt, return_tensors="pt").to(0)
+output = model.generate(inputs["input_ids"], max_new_tokens=128, do_sample=True, temperature=1.0, top_p=0.3, top_k=0, )
+print(tokenizer.decode(output[0].tolist(), skip_special_tokens=True))
+```
+
+output:
+
+```shell
+User: hi
+
+Assistant: Hi. I am your assistant and I will provide expert full response in full details. Please feel free to ask any question and I will always answer it.
+
+User: 介绍一下大熊猫
+
+Assistant: 大熊猫是一种中国特有的哺乳动物，也是中国的国宝之一。它们的外貌特征是圆形的黑白相间的身体，有着黑色的毛发和白色的耳朵。大熊猫的食物主要是竹子，它们会在竹林中寻找竹子，并且会将竹子放在竹笼中进行储存。大熊猫的寿命约为20至30年，但由于栖息地的丧失和人类活动的
+```
+
+#### Batch Inference
+
+```python
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+def generate_prompt(instruction, input=""):
+    instruction = instruction.strip().replace('\r\n', '\n').replace('\n\n', '\n')
+    input = input.strip().replace('\r\n', '\n').replace('\n\n', '\n')
+    if input:
+        return f"""Instruction: {instruction}
+
+Input: {input}
+
+Response:"""
+    else:
+        return f"""User: hi
+
+Assistant: Hi. I am your assistant and I will provide expert full response in full details. Please feel free to ask any question and I will always answer it.
+
+User: {instruction}
+
+Assistant:"""
+
+model = AutoModelForCausalLM.from_pretrained("RWKV/rwkv-5-world-1b5", trust_remote_code=True).to(torch.float32)
+tokenizer = AutoTokenizer.from_pretrained("RWKV/rwkv-5-world-1b5", trust_remote_code=True, padding_side='left', pad_token="<s>")
+
+texts = ["请介绍北京的旅游景点", "介绍一下大熊猫", "乌兰察布"]
+prompts = [generate_prompt(text) for text in texts]
+
+inputs = tokenizer(prompts, return_tensors="pt", padding=True)
+outputs = model.generate(inputs["input_ids"], max_new_tokens=128, do_sample=True, temperature=1.0, top_p=0.3, top_k=0, )
+
+for output in outputs:
+    print(tokenizer.decode(output.tolist(), skip_special_tokens=True))
+
+```
+
+output:
+
+```shell
+User: hi
+
+Assistant: Hi. I am your assistant and I will provide expert full response in full details. Please feel free to ask any question and I will always answer it.
+
+User: 请介绍北京的旅游景点
+
+Assistant: 北京是中国的首都，拥有丰富的旅游资源和历史文化遗产。以下是一些北京的旅游景点：
+1. 故宫：位于北京市中心，是明清两代的皇宫，是中国最大的古代宫殿建筑群之一。
+2. 天安门广场：位于北京市中心，是中国最著名的城市广场之一，也是中国最大的城市广场。
+3. 颐和
+User: hi
+
+Assistant: Hi. I am your assistant and I will provide expert full response in full details. Please feel free to ask any question and I will always answer it.
+
+User: 介绍一下大熊猫
+
+Assistant: 大熊猫是一种生活在中国中部地区的哺乳动物，也是中国的国宝之一。它们的外貌特征是圆形的黑白相间的身体，有着黑色的毛发和圆圆的眼睛。大熊猫是一种濒危物种，目前只有在野外的几个保护区才能看到它们的身影。大熊猫的食物主要是竹子，它们会在竹子上寻找食物，并且可以通
+User: hi
+
+Assistant: Hi. I am your assistant and I will provide expert full response in full details. Please feel free to ask any question and I will always answer it.
+
+User: 乌兰察布
+
+Assistant: 乌兰察布是中国新疆维吾尔自治区的一个县级市，位于新疆维吾尔自治区中部，是新疆的第二大城市。乌兰察布市是新疆的第一大城市，也是新疆的重要城市之一。乌兰察布市是新疆的经济中心，也是新疆的重要交通枢纽之一。乌兰察布市的人口约为2.5万人，其中汉族占绝大多数。乌
+```

added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<s>": 0
+}

config.json

@@ -0,0 +1,24 @@
+{
+  "architectures": [
+    "Rwkv5ForCausalLM"
+  ],
+  "auto_map": {
+      "AutoConfig": "configuration_rwkv5.Rwkv5Config",
+      "AutoModelForCausalLM": "modeling_rwkv5.Rwkv5ForCausalLM"
+  },
+  "attention_hidden_size": 2048,
+  "bos_token_id": 0,
+  "eos_token_id": 0,
+  "head_size": 64,
+  "hidden_size": 2048,
+  "intermediate_size": null,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "rwkv5",
+  "num_attention_heads": 64,
+  "num_hidden_layers": 24,
+  "rescale_every": 6,
+  "tie_word_embeddings": false,
+  "transformers_version": "4.34.0",
+  "use_cache": true,
+  "vocab_size": 65536
+}

configuration_rwkv5.py

@@ -0,0 +1,118 @@
+# coding=utf-8
+# Copyright 2023 The OpenAI Team Authors and HuggingFace Inc. team.
+# Copyright (c) 2018, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+""" RWKV configuration"""
+
+from transformers.configuration_utils import PretrainedConfig
+from transformers.utils import logging
+
+
+logger = logging.get_logger(__name__)
+
+RWKV5_PRETRAINED_CONFIG_ARCHIVE_MAP = {}
+
+
+class Rwkv5Config(PretrainedConfig):
+    """
+    This is the configuration class to store the configuration of a [`Rwkv5Model`]. It is used to instantiate a RWKV5
+    model according to the specified arguments, defining the model architecture. Instantiating a configuration with the
+    defaults will yield a similar configuration to that of the RWVK-4
+    [RWKV/rwkv-5-world-1b5](https://huggingface.co/RWKV/rwkv-5-world-1b5) architecture.
+
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+
+
+    Args:
+        vocab_size (`int`, *optional*, defaults to 65536):
+            Vocabulary size of the RWKV5 model. Defines the number of different tokens that can be represented by the
+            `inputs_ids` passed when calling [`Rwkv5Model`].
+        hidden_size (`int`, *optional*, defaults to 768):
+            Dimensionality of the embeddings and hidden states.
+        num_hidden_layers (`int`, *optional*, defaults to 24):
+            Number of hidden layers in the model.
+        attention_hidden_size (`int`, *optional*):
+            Dimensionality of the attention hidden states. Will default to `hidden_size` if unset.
+        num_attention_heads (`int`, *optional*, defaults to 64):
+            The attention heads to use in rwkv5 self_attention module.
+        head_size (`int`, *optional*, defaults to 64): head_size of rwkv5 self_attention module.
+        intermediate_size (`int`, *optional*):
+            Dimensionality of the inner feed-forward layers. Will default to 4 times `hidden_size` if unset.
+        layer_norm_epsilon (`float`, *optional*, defaults to 1e-05):
+            The epsilon to use in the layer normalization layers.
+        bos_token_id (`int`, *optional*, defaults to 0):
+            The id of the beginning of sentence token in the vocabulary. Defaults to 0.
+        eos_token_id (`int`, *optional*, defaults to 0):
+            The id of the end of sentence token in the vocabulary. Defaults to 0.
+        rescale_every (`int`, *optional*, defaults to 6):
+            At inference, the hidden states (and weights of the correponding output layers) are divided by 2 every
+            `rescale_every` layer. If set to 0 or a negative number, no rescale is done.
+        tie_word_embeddings (`bool`, *optional*, defaults to `False`):
+            Whether or not to tie the word embeddings with the input token embeddings.
+        use_cache (`bool`, *optional*, defaults to `True`):
+            Whether or not the model should return the last state.
+
+
+    Example:
+
+    ```python
+    >>> from transformers import Rwkv5Config, Rwkv5Model
+
+    >>> # Initializing a Rwkv5 configuration
+    >>> configuration = Rwkv5Config()
+
+    >>> # Initializing a model (with random weights) from the configuration
+    >>> model = Rwkv5Model(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "rwkv5"
+
+    def __init__(
+        self,
+        vocab_size=65536,
+        hidden_size=768,
+        num_hidden_layers=24,
+        attention_hidden_size=None,
+        head_size=64,
+        head_size_divisor=8,
+        intermediate_size=None,
+        layer_norm_epsilon=1e-5,
+        bos_token_id=0,
+        eos_token_id=0,
+        rescale_every=6,
+        tie_word_embeddings=False,
+        use_cache=True,
+        **kwargs,
+    ):
+        self.vocab_size = vocab_size
+        self.hidden_size = hidden_size
+        self.num_hidden_layers = num_hidden_layers
+        self.attention_hidden_size = attention_hidden_size if attention_hidden_size is not None else hidden_size
+        self.head_size = head_size
+        self.head_size_divisor = head_size_divisor
+        self.intermediate_size = None
+        self.layer_norm_epsilon = layer_norm_epsilon
+        self.rescale_every = rescale_every
+        self.use_cache = use_cache
+
+        self.bos_token_id = bos_token_id
+        self.eos_token_id = eos_token_id
+
+        super().__init__(
+            tie_word_embeddings=tie_word_embeddings, bos_token_id=bos_token_id, eos_token_id=eos_token_id, **kwargs
+        )

generation_config.json

@@ -0,0 +1,12 @@
+{
+    "chat_format": "chatml",
+    "eos_token_id": 0,
+    "pad_token_id": 0,
+    "max_window_size": 4096,
+    "max_new_tokens": 4096,
+    "do_sample": true,
+    "top_k": 0,
+    "top_p": 0.1,
+    "repetition_penalty": 1.0,
+    "transformers_version": "4.31.1"
+}

modeling_rwkv5.py

@@ -0,0 +1,704 @@
+# coding=utf-8
+# Copyright 2024 The RWKV team and HuggingFace Inc. team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""PyTorch RWKV5 World model."""
+
+from dataclasses import dataclass
+from pathlib import Path
+from typing import List, Optional, Tuple, Union
+
+import pkg_resources
+import torch
+import torch.nn.functional as F
+import torch.utils.checkpoint
+from torch import nn
+from torch.nn import CrossEntropyLoss
+
+from transformers.modeling_utils import PreTrainedModel
+from transformers.utils import (
+    ModelOutput,
+    add_code_sample_docstrings,
+    add_start_docstrings,
+    add_start_docstrings_to_model_forward,
+    is_bitsandbytes_available,
+    is_ninja_available,
+    is_torch_cuda_available,
+    logging,
+)
+
+try:
+    from flash_rwkv import rwkv5_cuda_linear_attention
+    # Check version
+    required_version = pkg_resources.parse_version("0.2.1")
+    current_version = pkg_resources.get_distribution("flash-rwkv").parsed_version
+
+    if current_version < required_version:
+        raise Exception("Your version of flash-rwkv is below 0.2.1. Please use pip install --upgrade flash-rwkv to update or install the required version.")
+except ImportError:
+    raise ImportError("The flash-rwkv package is not detected. Please install it using pip install flash-rwkv.")
+except pkg_resources.DistributionNotFound:
+    raise ImportError("The flash-rwkv package is not detected. Please install it using pip install flash-rwkv.")
+
+from .configuration_rwkv5 import Rwkv5Config
+
+
+logger = logging.get_logger(__name__)
+
+_CHECKPOINT_FOR_DOC = "RWKV/rwkv-5-world-1b5"
+_CONFIG_FOR_DOC = "Rwkv5Config"
+
+
+def rwkv5_linear_attention_cpu(receptance, key, value, time_decay, time_first, state):
+    input_dtype = receptance.dtype
+    # For CPU fallback. Will be slower and probably take more memory than the custom CUDA kernel if not executed
+    # within a torch.no_grad.
+    batch, seq_length, hidden_size = receptance.shape
+    num_heads, head_size = time_first.shape
+    key = key.float().view(batch, seq_length, num_heads, head_size).transpose(1, 2).transpose(-2, -1)
+    value = value.float().view(batch, seq_length, num_heads, head_size).transpose(1, 2)
+    receptance = receptance.float().view(batch, seq_length, num_heads, head_size).transpose(1, 2)
+    time_decay = torch.exp(-torch.exp(time_decay.float())).reshape(-1, 1, 1).reshape(num_heads, -1, 1)
+    time_first = time_first.float().reshape(-1, 1, 1).reshape(num_heads, -1, 1)
+    out = torch.zeros_like(key).reshape(batch, seq_length, num_heads, head_size)
+
+    for current_index in range(seq_length):
+        current_receptance = receptance[:, :, current_index:current_index+1, :]
+        current_key = key[:, :, :, current_index:current_index+1]
+        current_value = value[:, :, current_index:current_index+1, :]
+        attention_output = current_key @ current_value
+        out[:, current_index] = (current_receptance @ (time_first * attention_output + state)).squeeze(2)
+        with torch.no_grad():
+            state = attention_output + time_decay * state
+
+    return out, state
+
+# copied from RWKV but with receptance
+def RWKV5_linear_attention(training, receptance, key, value, time_decay, time_first, state):
+    no_cuda = any(t.device.type != "cuda" for t in [time_decay, time_first, key, value])
+    # Launching the CUDA kernel for just one token will actually be slower (there is no for loop in the CPU version
+    # in this case).
+    one_token = key.size(1) == 1
+    if not training or no_cuda or one_token:
+        return rwkv5_linear_attention_cpu(
+            receptance, key, value, time_decay, time_first, state
+        )
+    else:
+        return rwkv5_cuda_linear_attention(receptance.float(), key.float(), value.float(), time_decay.float().flatten(), time_first.float().flatten(), state)
+
+
+class Rwkv5SelfAttention(nn.Module):
+    def __init__(self, config, layer_id=0):
+        super().__init__()
+        self.config = config
+        self.layer_id = layer_id
+        hidden_size = config.hidden_size
+        attention_hidden_size = config.attention_hidden_size
+        self.attention_hidden_size = attention_hidden_size
+        head_size = config.head_size
+        num_heads = attention_hidden_size // head_size
+
+        self.time_decay = nn.Parameter(torch.empty(num_heads, head_size))
+        self.time_faaaa = nn.Parameter(torch.empty(num_heads, head_size))
+        self.time_mix_gate = nn.Parameter(torch.empty(1, 1, hidden_size))
+
+        self.time_mix_key = nn.Parameter(torch.empty(1, 1, hidden_size))
+        self.time_mix_value = nn.Parameter(torch.empty(1, 1, hidden_size))
+        self.time_mix_receptance = nn.Parameter(torch.empty(1, 1, hidden_size))
+
+        self.time_shift = nn.ZeroPad2d((0, 0, 1, -1))
+        self.key = nn.Linear(hidden_size, attention_hidden_size, bias=False)
+        self.value = nn.Linear(hidden_size, attention_hidden_size, bias=False)
+        self.receptance = nn.Linear(hidden_size, attention_hidden_size, bias=False)
+        self.gate = nn.Linear(hidden_size, attention_hidden_size, bias=False)
+        self.output = nn.Linear(attention_hidden_size, hidden_size, bias=False)
+        self.ln_x = nn.GroupNorm(num_heads, hidden_size)
+
+    def extract_key_value(self, hidden, state=None):
+        # Mix hidden with the previous timestep to produce key, value, receptance
+        if hidden.size(1) == 1 and state is not None:
+            shifted = state[0][:, :, self.layer_id]
+        else:
+            shifted = self.time_shift(hidden)
+            if state is not None:
+                shifted[:, 0] = state[0][:, :, self.layer_id]
+        if len(shifted.size()) == 2:
+            shifted = shifted.unsqueeze(1)
+        
+        key = hidden * self.time_mix_key + shifted * (1 - self.time_mix_key)
+        value = hidden * self.time_mix_value + shifted * (1 - self.time_mix_value)
+        receptance = hidden * self.time_mix_receptance + shifted * (1 - self.time_mix_receptance)
+        gate = hidden * self.time_mix_gate + shifted * (1 - self.time_mix_gate)
+
+        key = self.key(key)
+        value = self.value(value)
+        receptance = self.receptance(receptance)
+        gate = F.silu(self.gate(gate))
+
+        if state is not None:
+            state[0][:, :, self.layer_id] = hidden[:, -1]
+
+        return receptance, key, value, gate, state
+
+    def forward(self, hidden, state=None, use_cache=False, seq_mode=True):
+        receptance, key, value, gate, state = self.extract_key_value(hidden, state=state)
+
+        B,T,C = receptance.shape
+        H, S = self.time_faaaa.shape
+
+        layer_state = state[1][:, :, :, :, self.layer_id] if state is not None else None
+        out, layer_state = RWKV5_linear_attention(
+            self.training, receptance, key, value, self.time_decay, self.time_faaaa, layer_state
+        )
+
+        if layer_state is not None:
+            state[1][:, :, :, :, self.layer_id] = layer_state
+
+        out = out.reshape(B * T, H * S)
+        out = F.group_norm(out / self.config.head_size_divisor, num_groups=H, weight=self.ln_x.weight.to(out.dtype), bias=self.ln_x.bias.to(out.dtype), eps=self.ln_x.eps).reshape(B, T, H * S)
+        out = out.to(dtype=hidden.dtype) * gate
+        out = self.output(out)
+        return out, state
+
+# Copied from rwkv except for the intermediate size
+class Rwkv5FeedForward(nn.Module):
+    def __init__(self, config, layer_id=0):
+        super().__init__()
+        self.config = config
+        self.layer_id = layer_id
+        hidden_size = config.hidden_size
+        intermediate_size = (
+            config.intermediate_size
+            if config.intermediate_size is not None
+            else int((config.hidden_size * 3.5) // 32 * 32)
+        )
+
+        self.time_shift = nn.ZeroPad2d((0, 0, 1, -1))
+        self.time_mix_key = nn.Parameter(torch.empty(1, 1, hidden_size))
+        self.time_mix_receptance = nn.Parameter(torch.empty(1, 1, hidden_size))
+
+        self.key = nn.Linear(hidden_size, intermediate_size, bias=False)
+        self.receptance = nn.Linear(hidden_size, hidden_size, bias=False)
+        self.value = nn.Linear(intermediate_size, hidden_size, bias=False)
+
+    def forward(self, hidden, state=None):
+        if hidden.size(1) == 1 and state is not None:
+            shifted = state[2][:, :, self.layer_id]
+        else:
+            shifted = self.time_shift(hidden)
+            if state is not None:
+                shifted[:, 0] = state[2][:, :, self.layer_id]
+        if len(shifted.size()) == 2:
+            shifted = shifted.unsqueeze(1)
+        key = hidden * self.time_mix_key + shifted * (1 - self.time_mix_key)
+        receptance = hidden * self.time_mix_receptance + shifted * (1 - self.time_mix_receptance)
+
+        key = torch.square(torch.relu(self.key(key)))
+        value = self.value(key)
+        receptance = torch.sigmoid(self.receptance(receptance))
+
+        if state is not None:
+            state[2][:, :, self.layer_id] = hidden[:, -1]
+
+        return receptance * value, state
+
+
+# Copied from transformers.models.rwkv.modeling_rwkv.RwkvBlock with Rwkv->Rwkv5
+class Rwkv5Block(nn.Module):
+    def __init__(self, config, layer_id):
+        super().__init__()
+        self.config = config
+        self.layer_id = layer_id
+
+        if layer_id == 0:
+            self.pre_ln = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_epsilon)
+
+        self.ln1 = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_epsilon)
+        self.ln2 = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_epsilon)
+
+        self.attention = Rwkv5SelfAttention(config, layer_id)
+        self.feed_forward = Rwkv5FeedForward(config, layer_id)
+
+    def forward(self, hidden, state=None, use_cache=False, output_attentions=False, seq_mode=True):
+        if self.layer_id == 0:
+            hidden = self.pre_ln(hidden)
+        attention, state = self.attention(self.ln1(hidden), state=state, use_cache=use_cache, seq_mode=seq_mode)
+        hidden = hidden + attention
+
+        feed_forward, state = self.feed_forward(self.ln2(hidden), state=state)
+        hidden = hidden + feed_forward
+
+        outputs = (hidden, state)
+        if output_attentions:
+            outputs += (attention,)
+        else:
+            outputs += (None,)
+
+        return outputs
+
+
+# Copied from transformers.models.rwkv.modeling_rwkv.RwkvPreTrainedModel with Rwkv->Rwkv5
+class Rwkv5PreTrainedModel(PreTrainedModel):
+    """
+    An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained
+    models.
+    """
+
+    config_class = Rwkv5Config
+    base_model_prefix = "rwkv5"
+    _no_split_modules = ["Rwkv5Block"]
+    _keep_in_fp32_modules = ["time_decay", "time_first"]
+    supports_gradient_checkpointing = True
+
+    def _init_weights(self, module):
+        """Initialize the weights."""
+        if isinstance(module, Rwkv5SelfAttention):
+            layer_id = module.layer_id
+            num_hidden_layers = module.config.num_hidden_layers
+            hidden_size = module.config.hidden_size
+            attention_hidden_size = module.attention_hidden_size
+            head_size = module.config.head_size
+            num_heads = attention_hidden_size // head_size
+
+            ratio_0_to_1 = layer_id / (num_hidden_layers - 1)  # 0 to 1
+            ratio_1_to_almost0 = 1.0 - (layer_id / num_hidden_layers)  # 1 to ~0
+
+            time_weight = torch.tensor(
+                [i / hidden_size for i in range(hidden_size)],
+                dtype=module.time_mix_key.dtype,
+                device=module.time_mix_key.device,
+            )
+            time_weight = time_weight[None, None, :]
+
+            decay_speed = [
+                -6.0 + 5.0 * (h / (attention_hidden_size - 1)) ** (0.7 + 1.3 * ratio_0_to_1)
+                for h in range(attention_hidden_size)
+            ]
+            decay_speed = torch.tensor(decay_speed, dtype=module.time_decay.dtype, device=module.time_decay.device)
+            tmp = torch.tensor(
+                [
+                    (1.0 - (i / (attention_hidden_size - 1.0))) * ratio_0_to_1 + 0.1 * ((i + 1) % 3 - 1)
+                    for i in range(attention_hidden_size)
+                ],
+                dtype=module.time_faaaa.dtype,
+                device=module.time_faaaa.device,
+            )
+
+            with torch.no_grad():
+                module.time_decay.data = decay_speed.reshape(num_heads, head_size)
+                module.time_faaaa.data = tmp.reshape(num_heads, head_size)
+                module.time_mix_key.data = torch.pow(time_weight, ratio_1_to_almost0)
+
+                module.time_mix_value.data = torch.pow(time_weight, ratio_1_to_almost0) + 0.3 * ratio_0_to_1
+                module.time_mix_receptance.data = torch.pow(time_weight, 0.5 * ratio_1_to_almost0)
+                module.time_mix_gate.data = torch.pow(time_weight, 0.5 * ratio_1_to_almost0)
+
+        elif isinstance(module, Rwkv5FeedForward):
+            layer_id = module.layer_id
+            num_hidden_layers = module.config.num_hidden_layers
+            hidden_size = module.config.hidden_size
+
+            ratio_1_to_almost0 = 1.0 - (layer_id / num_hidden_layers)  # 1 to ~0
+
+            time_weight = torch.tensor(
+                [i / hidden_size for i in range(hidden_size)],
+                dtype=module.time_mix_key.dtype,
+                device=module.time_mix_key.device,
+            )
+            time_weight = time_weight[None, None, :]
+
+            with torch.no_grad():
+                module.time_mix_key.data = torch.pow(time_weight, ratio_1_to_almost0)
+                module.time_mix_receptance.data = torch.pow(time_weight, ratio_1_to_almost0)
+
+
+# Copied from transformers.models.rwkv.modeling_rwkv.RwkvOutput with Rwkv->Rwkv5
+@dataclass
+class Rwkv5Output(ModelOutput):
+    """
+    Class for the RWKV5 model outputs.
+
+    Args:
+        last_hidden_state (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`):
+            Sequence of hidden-states at the output of the last layer of the model.
+        state (list of five `torch.FloatTensor` of shape `(batch_size, hidden_size, num_hidden_layers)`):
+            The state of the model at the last time step. Can be used in a forward method with the next `input_ids` to
+            avoid providing the old `input_ids`.
+        hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`):
+            Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
+            one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`. Hidden-states of
+            the model at the output of each layer plus the optional initial embedding outputs.
+        attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`):
+            Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
+            sequence_length)`. Attentions weights after the attention softmax, used to compute the weighted average in
+            the self-attention heads.
+    """
+
+    last_hidden_state: torch.FloatTensor = None
+    state: Optional[List[torch.FloatTensor]] = None
+    hidden_states: Optional[Tuple[torch.FloatTensor]] = None
+    attentions: Optional[Tuple[torch.FloatTensor]] = None
+
+
+# Copied from transformers.models.rwkv.modeling_rwkv.RwkvCausalLMOutput with Rwkv->Rwkv5
+@dataclass
+class Rwkv5CausalLMOutput(ModelOutput):
+    """
+    Base class for causal language model (or autoregressive) outputs.
+
+    Args:
+        loss (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `labels` is provided):
+            Language modeling loss (for next-token prediction).
+        logits (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.vocab_size)`):
+            Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
+        state (list of five `torch.FloatTensor` of shape `(batch_size, hidden_size, num_hidden_layers)`):
+            The state of the model at the last time step. Can be used in a forward method with the next `input_ids` to
+            avoid providing the old `input_ids`.
+        hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`):
+            Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
+            one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`. Hidden-states of
+            the model at the output of each layer plus the optional initial embedding outputs.
+        attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`):
+            Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
+            sequence_length)`. Attentions weights after the attention softmax, used to compute the weighted average in
+            the self-attention heads.
+    """
+
+    loss: Optional[torch.FloatTensor] = None
+    logits: torch.FloatTensor = None
+    state: Optional[List[torch.FloatTensor]] = None
+    hidden_states: Optional[Tuple[torch.FloatTensor]] = None
+    attentions: Optional[Tuple[torch.FloatTensor]] = None
+
+
+RWKV5_START_DOCSTRING = r"""
+    This model inherits from [`PreTrainedModel`]. Check the superclass documentation for the generic methods the
+    library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads
+    etc.) This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module)
+    subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to
+    general usage and behavior.
+
+    Parameters:
+        config ([`Rwkv5Config`]): Model configuration class with all the parameters of the model.
+            Initializing with a config file does not load the weights associated with the model, only the
+            configuration. Check out the [`~PreTrainedModel.from_pretrained`] method to load the model weights.
+"""
+
+RWKV5_INPUTS_DOCSTRING = r"""
+    Args:
+        input_ids (`torch.LongTensor` of shape `(batch_size, input_ids_length)`):
+            `input_ids_length` = `sequence_length` if `past_key_values` is `None` else
+            `past_key_values[0][0].shape[-2]` (`sequence_length` of input past key value states). Indices of input
+            sequence tokens in the vocabulary. If `past_key_values` is used, only `input_ids` that do not have their
+            past calculated should be passed as `input_ids`. Indices can be obtained using [`AutoTokenizer`]. See
+            [`PreTrainedTokenizer.encode`] and [`PreTrainedTokenizer.__call__`] for details. [What are input
+            IDs?](../glossary#input-ids)
+        inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
+            Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This
+            is useful if you want more control over how to convert `input_ids` indices into associated vectors than the
+            model's internal embedding lookup matrix.
+        state (tuple of five `torch.FloatTensor` of shape `(batch_size, hidden_size, num_hidden_layers)`, *optional*):
+            If passed along, the model uses the previous state in all the blocks (which will give the output for the
+            `input_ids` provided as if the model add `state_input_ids + input_ids` as context).
+        use_cache (`bool`, *optional*):
+            If set to `True`, the last state is returned and can be used to quickly generate the next logits.
+        output_attentions (`bool`, *optional*):
+            Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
+            tensors for more detail.
+        output_hidden_states (`bool`, *optional*):
+            Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
+            more detail.
+        return_dict (`bool`, *optional*):
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
+"""
+
+
+@add_start_docstrings(
+    "The bare RWKV5 Model transformer outputting raw hidden-states without any specific head on top.",
+    RWKV5_START_DOCSTRING,
+)
+class Rwkv5Model(Rwkv5PreTrainedModel):
+    def __init__(self, config):
+        super().__init__(config)
+
+        self.embeddings = nn.Embedding(config.vocab_size, config.hidden_size)
+        self.blocks = nn.ModuleList([Rwkv5Block(config, layer_id=idx) for idx in range(config.num_hidden_layers)])
+        self.ln_out = nn.LayerNorm(config.hidden_size)
+
+        self.layers_are_rescaled = False
+        self.gradient_checkpointing = False
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def get_input_embeddings(self):
+        return self.embeddings
+
+    def set_input_embeddings(self, new_embeddings):
+        self.embeddings = new_embeddings
+
+    @add_start_docstrings_to_model_forward(RWKV5_INPUTS_DOCSTRING)
+    @add_code_sample_docstrings(
+        checkpoint=_CHECKPOINT_FOR_DOC,
+        output_type=Rwkv5Output,
+        config_class=_CONFIG_FOR_DOC,
+    )
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        attention_mask: Optional[torch.LongTensor] = None,  # noqa
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        state: Optional[List[torch.FloatTensor]] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, Rwkv5Output]:
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        # FIXME - training is supportable with the CUDA code
+        # rwkv5 only support inference in huggingface.
+        use_cache = use_cache if use_cache is not None else self.config.use_cache
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        if self.training == self.layers_are_rescaled and (
+            self.embeddings.weight.dtype == torch.float16 or self.embeddings.weight.dtype == torch.bfloat16
+        ):
+            self._rescale_layers()
+
+        if input_ids is not None and inputs_embeds is not None:
+            raise ValueError("You cannot specify both input_ids and inputs_embeds at the same time")
+        elif input_ids is None and inputs_embeds is None:
+            raise ValueError("You have to specify either input_ids or inputs_embeds")
+
+        if inputs_embeds is None:
+            inputs_embeds = self.embeddings(input_ids)
+
+        if state is None:
+            state = []
+            head_size = self.config.head_size
+            num_heads = self.config.attention_hidden_size // head_size
+            state_attn_x = torch.zeros(
+                (inputs_embeds.size(0), self.config.hidden_size, self.config.num_hidden_layers),
+                dtype=inputs_embeds.dtype,
+                requires_grad=False,
+                device=inputs_embeds.device,
+            ).contiguous()
+            state_attn_kv = torch.zeros(
+                (
+                    inputs_embeds.size(0),
+                    num_heads,
+                    head_size,
+                    head_size,
+                    self.config.num_hidden_layers,
+                ),
+                dtype=torch.float32,
+                requires_grad=False,
+                device=inputs_embeds.device,
+            ).contiguous()
+            state_ffn_x = torch.zeros(
+                (inputs_embeds.size(0), self.config.hidden_size, self.config.num_hidden_layers),
+                dtype=inputs_embeds.dtype,
+                requires_grad=False,
+                device=inputs_embeds.device,
+            ).contiguous()
+            state.append(state_attn_x)
+            state.append(state_attn_kv)
+            state.append(state_ffn_x)
+
+        seq_mode = inputs_embeds.shape[1] > 1
+        hidden_states = inputs_embeds
+
+        all_self_attentions = () if output_attentions else None
+        all_hidden_states = () if output_hidden_states else None
+        for idx, block in enumerate(self.blocks):
+            hidden_states, state, attentions = block(
+                hidden_states, state=state, use_cache=use_cache, output_attentions=output_attentions, seq_mode=seq_mode
+            )
+            if (
+                self.layers_are_rescaled
+                and self.config.rescale_every > 0
+                and (idx + 1) % self.config.rescale_every == 0
+            ):
+                hidden_states = hidden_states / 2
+
+            if output_hidden_states:
+                all_hidden_states = all_hidden_states + (hidden_states,)
+
+            if output_attentions:
+                all_self_attentions = all_self_attentions + (attentions,)
+
+        hidden_states = self.ln_out(hidden_states)
+
+        if output_hidden_states:
+            all_hidden_states = all_hidden_states + (hidden_states,)
+
+        if not return_dict:
+            return (hidden_states, state, all_hidden_states, all_self_attentions)
+
+        return Rwkv5Output(
+            last_hidden_state=hidden_states,
+            state=state,
+            hidden_states=all_hidden_states,  # None
+            attentions=all_self_attentions,  # None
+        )
+
+    def _rescale_layers(self):
+        # Layers should be rescaled for inference only.
+        if self.layers_are_rescaled == (not self.training):
+            return
+        if self.config.rescale_every > 0:
+            with torch.no_grad():
+                for block_id, block in enumerate(self.blocks):
+                    if self.training:
+                        block.attention.output.weight.mul_(2 ** int(block_id // self.config.rescale_every))
+                        block.feed_forward.value.weight.mul_(2 ** int(block_id // self.config.rescale_every))
+                    else:
+                        # Deal with quantization statistics
+                        if hasattr(block.attention.output.weight, "SCB"):
+                            block.attention.output.weight.SCB.div_(2 ** int(block_id // self.config.rescale_every))
+                            block.feed_forward.value.weight.SCB.div_(2 ** int(block_id // self.config.rescale_every))
+                        elif hasattr(block.attention.output.weight, "quant_state"):
+                            self._bnb_4bit_dequantize_and_rescale(block.attention.output, block_id)
+                            self._bnb_4bit_dequantize_and_rescale(block.feed_forward.value, block_id)
+                        else:
+                            block.attention.output.weight.div_(2 ** int(block_id // self.config.rescale_every))
+                            block.feed_forward.value.weight.div_(2 ** int(block_id // self.config.rescale_every))
+
+        self.layers_are_rescaled = not self.training
+
+    def _bnb_4bit_dequantize_and_rescale(self, target_layer, block_id):
+        r"""
+        Perform the dequantization and rescaling of the weights of a given layer. After that operation the layer will
+        be quantized again.
+        """
+        if not is_bitsandbytes_available():
+            raise ImportError("Please install bitsandbytes to use this method.")
+        import bitsandbytes as bnb
+
+        dequant_weights = bnb.functional.dequantize_4bit(target_layer.weight.data, target_layer.weight.quant_state)
+
+        dequant_weights.div_(2 ** int(block_id // self.config.rescale_every))
+
+        # re-quantize the model:
+        # we need to put it first on CPU then back to the device
+        # this will create an overhead :/
+        # We set requires_grad=False as we cannot compute gradients on top of 4bit parameters anyway and to avoid
+        # bugs with bnb
+        quant_weight = bnb.nn.Params4bit(dequant_weights.to("cpu"), requires_grad=False).to(dequant_weights.device)
+        setattr(target_layer, "weight", quant_weight)
+
+
+# copied from HuggingFace https://github.com/huggingface/transformers/blob/main/src/transformers/models/rwkv/modeling_rwkv.py
+@add_start_docstrings(
+    """
+    The RWKV5 Model transformer with a language modeling head on top (linear layer with weights tied to the input
+    embeddings).
+    """,
+    RWKV5_START_DOCSTRING,
+)
+# Copied from transformers.models.rwkv.modeling_rwkv.RwkvForCausalLM with Rwkv->Rwkv5
+class Rwkv5ForCausalLM(Rwkv5PreTrainedModel):
+    _tied_weights_keys = ["head.weight"]
+
+    def __init__(self, config):
+        super().__init__(config)
+        self.rwkv = Rwkv5Model(config)
+        self.head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def get_output_embeddings(self):
+        return self.head
+
+    def set_output_embeddings(self, new_embeddings):
+        self.head = new_embeddings
+
+    def prepare_inputs_for_generation(self, input_ids, state=None, inputs_embeds=None, **kwargs):
+        # only last token for inputs_ids if the state is passed along.
+        if state is not None:
+            input_ids = input_ids[:, -1].unsqueeze(-1)
+
+        # if `inputs_embeds` are passed, we only want to use them in the 1st generation step
+        if inputs_embeds is not None and state is None:
+            model_inputs = {"inputs_embeds": inputs_embeds}
+        else:
+            model_inputs = {"input_ids": input_ids}
+
+        model_inputs["state"] = state
+        return model_inputs
+
+    @add_start_docstrings_to_model_forward(RWKV5_INPUTS_DOCSTRING)
+    @add_code_sample_docstrings(
+        checkpoint=_CHECKPOINT_FOR_DOC,
+        output_type=Rwkv5CausalLMOutput,
+        config_class=_CONFIG_FOR_DOC,
+    )
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        attention_mask: Optional[torch.LongTensor] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        state: Optional[List[torch.FloatTensor]] = None,
+        labels: Optional[torch.LongTensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, Rwkv5CausalLMOutput]:
+        r"""
+        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Labels for language modeling. Note that the labels **are shifted** inside the model, i.e. you can set
+            `labels = input_ids` Indices are selected in `[-100, 0, ..., config.vocab_size]` All labels set to `-100`
+            are ignored (masked), the loss is only computed for labels in `[0, ..., config.vocab_size]`
+        """
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        outputs = self.rwkv(
+            input_ids,
+            inputs_embeds=inputs_embeds,
+            state=state,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+        hidden_states = outputs[0]
+
+        logits = self.head(hidden_states)
+
+        loss = None
+        if labels is not None:
+            # move labels to correct device to enable model parallelism
+            labels = labels.to(logits.device)
+            # Shift so that tokens < n predict n
+            shift_logits = logits[..., :-1, :].contiguous()
+            shift_labels = labels[..., 1:].contiguous()
+            # Flatten the tokens
+            loss_fct = CrossEntropyLoss()
+            loss = loss_fct(shift_logits.view(-1, shift_logits.size(-1)), shift_labels.view(-1))
+
+        if not return_dict:
+            output = (logits,) + outputs[1:]
+            return ((loss,) + output) if loss is not None else output
+
+        return Rwkv5CausalLMOutput(
+            loss=loss,
+            logits=logits,
+            state=outputs.state,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )

special_tokens_map.json

@@ -0,0 +1,5 @@
+{
+  "bos_token": "<s>",
+  "eos_token": "<s>",
+  "unk_token": "<s>"
+}

tokenization_rwkv5.py

@@ -0,0 +1,229 @@
+# coding=utf-8
+# Copyright 2024 The HuggingFace Inc. team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""Tokenization classes for RWKV5."""
+
+import os
+import re
+from typing import TYPE_CHECKING, List, Optional, Tuple
+
+from transformers.tokenization_utils import AddedToken, PreTrainedTokenizer
+from transformers.utils import logging
+
+
+if TYPE_CHECKING:
+    pass
+
+logger = logging.get_logger(__name__)
+
+VOCAB_FILES_NAMES = {
+    "vocab_file": "vocab.txt",
+}
+PRETRAINED_VOCAB_FILES_MAP = {
+    "vocab_file": {
+        "ArthurZ/rwkv-5-utf": "https://huggingface.co/ArthurZ/rwkv-5-utf/blob/main/vocab.txt",
+    },
+}
+
+
+def whitespace_tokenize(text):
+    """Runs basic whitespace cleaning and splitting on a piece of text.
+    The separators are kept
+    """
+    text = text.strip()
+    if not text:
+        return []
+    tokens = re.split(b"(?= )", text)
+    return tokens
+
+
+class WordpieceTokenizer(object):
+    """Runs WordPiece tokenization."""
+
+    def __init__(self, vocab, unk_token):
+        self.vocab = vocab
+        self.unk_token = unk_token
+
+    def tokenize(self, text):
+        """
+        Tokenizes a piece of text into its word pieces. This uses a greedy longest-match-first algorithm to perform
+        tokenization using the given vocabulary.
+
+        For example, `input = "unaffable"` wil return as output `["un", "##aff", "##able"]`.
+
+        Args:
+            text: A single token or whitespace separated tokens. This should have
+                already been passed through *BasicTokenizer*.
+
+        Returns:
+            A list of wordpiece tokens.
+        """
+
+        output_tokens = []
+        for token in whitespace_tokenize(text):
+            chars = list(token)
+            is_bad = False
+            start = 0
+            sub_tokens = []
+            while start < len(chars):
+                end = len(chars)
+                cur_substr = None
+                while start < end:
+                    substr = bytes(chars[start:end])
+                    if substr in self.vocab:
+                        cur_substr = substr
+                        break
+                    end -= 1
+                if cur_substr is None:
+                    is_bad = True
+                    break
+                try:
+                    cur_substr = cur_substr.decode()
+                except UnicodeDecodeError:
+                    cur_substr = str(cur_substr)
+                sub_tokens.append(cur_substr)
+                start = end
+            if is_bad:
+                output_tokens.append(self.unk_token)
+            else:
+                output_tokens.extend(sub_tokens)
+        return output_tokens
+
+
+class Rwkv5Tokenizer(PreTrainedTokenizer):
+    vocab_files_names = VOCAB_FILES_NAMES
+    pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
+    max_model_input_sizes = {"ArthurZ/rwkv-5-utf": 2048}
+
+    model_input_names = ["input_ids", "attention_mask"]
+
+    def __init__(self, vocab_file, bos_token="<s>", eos_token="<s>", unk_token="<s>", **kwargs):
+        if not os.path.isfile(vocab_file):
+            raise ValueError(
+                f"Can't find a vocabulary file at path '{vocab_file}'. To load the vocabulary from a Google pretrained"
+                " model use `tokenizer = BertTokenizer.from_pretrained(PRETRAINED_MODEL_NAME)`"
+            )
+
+        with open(vocab_file, "r") as reader:
+            tokens = reader.readlines()
+        vocab = {}
+        for index, token in enumerate(tokens):
+            token = eval(token.rstrip("\n"))
+            vocab[token] = index
+
+        self.add_bos_token = True
+        self.encoder = vocab
+        self.decoder = {v: k for k, v in vocab.items()}
+        self.wordpiece_tokenizer = WordpieceTokenizer(vocab=self.encoder, unk_token=str(unk_token))
+        self._added_tokens_decoder = {0: AddedToken(str(bos_token))}
+        super().__init__(bos_token=bos_token, eos_token=eos_token, unk_token=unk_token, **kwargs)
+
+    @property
+    def vocab_size(self):
+        return len(self.encoder)
+
+    def get_vocab(self):
+        vocab = {str(self.convert_ids_to_tokens(i)): i for i in range(self.vocab_size)}
+        vocab.update(self.added_tokens_encoder)
+        return vocab
+
+    def _tokenize(self, text, split_special_tokens=False):
+        return self.wordpiece_tokenizer.tokenize(text.encode("utf-8"))
+
+    def _convert_token_to_id(self, token):
+        """Converts a token (byte) to an id using the vocab."""
+        if token.startswith("b'\\"):
+            token = eval(token)
+        elif not isinstance(token, bytes):
+            token = token.encode("utf-8", errors="replace")
+        return self.encoder.get(token, self.unk_token_id)
+
+    def _convert_id_to_token(self, index):
+        """Converts an index (integer) in a token (byte) using the vocab."""
+        token = self.decoder.get(index, self.unk_token)
+        if isinstance(token, (bytes)):
+            token = token.decode("utf-8", errors="replace")
+        return token
+
+    def convert_tokens_to_string(self, tokens):
+        """Converts a sequence of tokens (bytes) in a single string. Additional tokens are encoded to bytes"""
+        out_string = b"".join([k.encode(errors="replace") if isinstance(k, str) else k for k in tokens]).decode(
+            "utf-8"
+        )
+        return out_string
+
+    def save_vocabulary(self, save_directory: str, filename_prefix: Optional[str] = None) -> Tuple[str]:
+        index = 0
+        if os.path.isdir(save_directory):
+            vocab_file = os.path.join(
+                save_directory, (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"]
+            )
+        else:
+            vocab_file = (filename_prefix + "-" if filename_prefix else "") + save_directory
+        with open(vocab_file, "w") as writer:
+            for token, token_index in sorted(self.encoder.items(), key=lambda kv: kv[1]):
+                if index != token_index:
+                    logger.warning(
+                        f"Saving vocabulary to {vocab_file}: vocabulary indices are not consecutive."
+                        " Please check that the vocabulary is not corrupted!"
+                    )
+                    index = token_index
+                writer.write(str(token) + "\n")
+                index += 1
+        return (vocab_file,)
+
+    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
+        if self.add_bos_token:
+            bos_token_ids = [self.bos_token_id]
+        else:
+            bos_token_ids = []
+
+        output = bos_token_ids + token_ids_0
+
+        if token_ids_1 is None:
+            return output
+
+        return output + bos_token_ids + token_ids_1
+
+    def get_special_tokens_mask(
+        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None, already_has_special_tokens: bool = False
+    ) -> List[int]:
+        """
+        Retrieves sequence ids from a token list that has no special tokens added. This method is called when adding
+        special tokens using the tokenizer `prepare_for_model` or `encode_plus` methods.
+
+        Args:
+            token_ids_0 (`List[int]`):
+                List of IDs.
+            token_ids_1 (`List[int]`, *optional*):
+                Optional second list of IDs for sequence pairs.
+            already_has_special_tokens (`bool`, *optional*, defaults to `False`):
+                Whether or not the token list is already formatted with special tokens for the model.
+
+        Returns:
+            `List[int]`: A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token.
+        """
+        if already_has_special_tokens:
+            return super().get_special_tokens_mask(
+                token_ids_0=token_ids_0, token_ids_1=token_ids_1, already_has_special_tokens=True
+            )
+
+        if not self.add_bos_token:
+            return super().get_special_tokens_mask(
+                token_ids_0=token_ids_0, token_ids_1=token_ids_1, already_has_special_tokens=False
+            )
+
+        if token_ids_1 is None:
+            return [1] + ([0] * len(token_ids_0))
+        return [1] + ([0] * len(token_ids_0)) + [1] + ([0] * len(token_ids_1))

tokenizer_config.json

@@ -0,0 +1,12 @@
+{
+  "name_or_path": "rwkv-5-tokenizer",
+  "add_prefix_space": false,
+  "tokenizer_class": "Rwkv5Tokenizer",
+  "use_fast": false,
+  "auto_map": {
+    "AutoTokenizer": [
+      "tokenization_rwkv5.Rwkv5Tokenizer",
+      null
+      ]
+  }
+}