初始化项目，由ModelHub XC社区提供模型
Model: arcee-ai/Trinity-Nano-Preview Source: Original Platform
This commit is contained in:
ModelHub XC
53
.gitattributes
vendored
Normal file
53
.gitattributes
vendored
Normal file
@@ -0,0 +1,53 @@
|
|||||||
|
*.7z filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.arrow filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.bin filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.bin.* filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.ftz filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.gz filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.h5 filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.joblib filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.model filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.onnx filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.ot filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.parquet filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.pb filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.pt filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.pth filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.rar filter=lfs diff=lfs merge=lfs -text
|
||||||
|
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.tflite filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.tgz filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.xz filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.zip filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.zstandard filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.tfevents* filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.db* filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.ark* filter=lfs diff=lfs merge=lfs -text
|
||||||
|
**/*ckpt*data* filter=lfs diff=lfs merge=lfs -text
|
||||||
|
**/*ckpt*.meta filter=lfs diff=lfs merge=lfs -text
|
||||||
|
**/*ckpt*.index filter=lfs diff=lfs merge=lfs -text
|
||||||
|
|
||||||
|
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.gguf* filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.ggml filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.llamafile* filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.pt2 filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.npy filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.npz filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.pickle filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.pkl filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.tar filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.wasm filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*.zst filter=lfs diff=lfs merge=lfs -text
|
||||||
|
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
||||||
|
|
||||||
|
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
||||||
|
model-00003-of-00003.safetensors filter=lfs diff=lfs merge=lfs -text
|
||||||
|
model-00001-of-00003.safetensors filter=lfs diff=lfs merge=lfs -text
|
||||||
|
model.safetensors.index.json filter=lfs diff=lfs merge=lfs -text
|
||||||
|
model-00002-of-00003.safetensors filter=lfs diff=lfs merge=lfs -text
|
||||||
49
LICENSE
Normal file
49
LICENSE
Normal file
@@ -0,0 +1,49 @@
|
|||||||
|
OpenMDW License Agreement, version 1.1 (OpenMDW-1.1)
|
||||||
|
|
||||||
|
By exercising rights granted to you under this agreement, you accept and agree
|
||||||
|
to its terms.
|
||||||
|
|
||||||
|
As used in this agreement, "Model Materials" means the materials provided to
|
||||||
|
you under this agreement, consisting of: (1) one or more machine learning
|
||||||
|
models (including architecture and parameters); and (2) all related artifacts
|
||||||
|
(including associated data, documentation and software) that are provided to
|
||||||
|
you hereunder.
|
||||||
|
|
||||||
|
Subject to your compliance with this agreement, permission is hereby granted,
|
||||||
|
free of charge, to deal in the Model Materials without restriction, including
|
||||||
|
under all copyright, patent, database, and trade secret rights included or
|
||||||
|
embodied therein.
|
||||||
|
|
||||||
|
If you distribute any portion of the Model Materials, you shall retain in your
|
||||||
|
distribution (1) a copy of this agreement, and (2) all copyright notices and
|
||||||
|
other notices of origin included in the Model Materials that are applicable to
|
||||||
|
your distribution.
|
||||||
|
|
||||||
|
If you file, maintain, or voluntarily participate in a lawsuit against any
|
||||||
|
person or entity asserting that the Model Materials directly or indirectly
|
||||||
|
infringe any patent or copyright, then all rights and grants made to you
|
||||||
|
hereunder are terminated, unless that lawsuit was in response to a
|
||||||
|
corresponding lawsuit first brought against you.
|
||||||
|
|
||||||
|
This agreement does not impose any restrictions or obligations with respect to
|
||||||
|
any use, modification, or sharing of any outputs generated by using the Model
|
||||||
|
Materials.
|
||||||
|
|
||||||
|
THE MODEL MATERIALS ARE PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
|
||||||
|
OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
||||||
|
FITNESS FOR A PARTICULAR PURPOSE, TITLE, NONINFRINGEMENT, ACCURACY, OR THE
|
||||||
|
ABSENCE OF LATENT OR OTHER DEFECTS OR ERRORS, WHETHER OR NOT DISCOVERABLE, ALL
|
||||||
|
TO THE GREATEST EXTENT PERMISSIBLE UNDER APPLICABLE LAW.
|
||||||
|
|
||||||
|
YOU ARE SOLELY RESPONSIBLE FOR (1) CLEARING RIGHTS OF OTHER PERSONS THAT MAY
|
||||||
|
APPLY TO THE MODEL MATERIALS OR ANY USE THEREOF, INCLUDING WITHOUT LIMITATION
|
||||||
|
ANY PERSON'S COPYRIGHTS OR OTHER RIGHTS INCLUDED OR EMBODIED IN THE MODEL
|
||||||
|
MATERIALS; (2) OBTAINING ANY NECESSARY CONSENTS, PERMISSIONS OR OTHER RIGHTS
|
||||||
|
REQUIRED FOR ANY USE OF THE MODEL MATERIALS; OR (3) PERFORMING ANY DUE
|
||||||
|
DILIGENCE OR UNDERTAKING ANY OTHER INVESTIGATIONS INTO THE MODEL MATERIALS OR
|
||||||
|
ANYTHING INCORPORATED OR EMBODIED THEREIN.
|
||||||
|
|
||||||
|
IN NO EVENT SHALL THE PROVIDERS OF THE MODEL MATERIALS BE LIABLE FOR ANY CLAIM,
|
||||||
|
DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
|
||||||
|
OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE MODEL MATERIALS, THE
|
||||||
|
USE THEREOF OR OTHER DEALINGS THEREIN.
|
||||||
181
README.md
Normal file
181
README.md
Normal file
@@ -0,0 +1,181 @@
|
|||||||
|
---
|
||||||
|
license: other
|
||||||
|
language:
|
||||||
|
- en
|
||||||
|
- es
|
||||||
|
- fr
|
||||||
|
- de
|
||||||
|
- it
|
||||||
|
- pt
|
||||||
|
- ru
|
||||||
|
- ar
|
||||||
|
- hi
|
||||||
|
- ko
|
||||||
|
- zh
|
||||||
|
library_name: transformers
|
||||||
|
base_model:
|
||||||
|
- arcee-ai/Trinity-Nano-Base
|
||||||
|
license_link: LICENSE
|
||||||
|
license_name: openmdw-1.1
|
||||||
|
---
|
||||||
|
<div align="center">
|
||||||
|
<picture>
|
||||||
|
<img
|
||||||
|
src="https://cdn-uploads.huggingface.co/production/uploads/6435718aaaef013d1aec3b8b/i-v1KyAMOW_mgVGeic9WJ.png"
|
||||||
|
alt="Arcee Trinity Mini"
|
||||||
|
style="max-width: 100%; height: auto;"
|
||||||
|
>
|
||||||
|
</picture>
|
||||||
|
</div>
|
||||||
|
|
||||||
|
# Trinity Nano Preview
|
||||||
|
|
||||||
|
Trinity Nano Preview is a preview of Arcee AI's 6B MoE model with 1B active parameters. It is the small-sized model in our new Trinity family, a series of open-weight models for enterprise and tinkerers alike.
|
||||||
|
|
||||||
|
This is a chat tuned model, with a delightful personality and charm we think users will love. We note that this model is pushing the limits of sparsity in small language models with only 800M non-embedding parameters active per token, and as such **may be unstable** in certain use cases, especially in this preview.
|
||||||
|
|
||||||
|
This is an *experimental* release, it's fun to talk to but will not be hosted anywhere, so download it and try it out yourself!
|
||||||
|
|
||||||
|
***
|
||||||
|
|
||||||
|
Trinity Nano Preview is trained on 10T tokens gathered and curated through a key partnership with [Datology](https://www.datologyai.com/), building upon the excellent dataset we used on [AFM-4.5B](https://huggingface.co/arcee-ai/AFM-4.5B) with additional math and code.
|
||||||
|
|
||||||
|
Training was performed on a cluster of 512 H200 GPUs powered by [Prime Intellect](https://www.primeintellect.ai/) using HSDP parallelism.
|
||||||
|
|
||||||
|
More details, including key architecture decisions, can be found on our blog [here](https://www.arcee.ai/blog/the-trinity-manifesto)
|
||||||
|
|
||||||
|
***
|
||||||
|
|
||||||
|
## Model Details
|
||||||
|
|
||||||
|
* **Model Architecture:** AfmoeForCausalLM
|
||||||
|
* **Parameters:** 6B, 1B active
|
||||||
|
* **Experts:** 128 total, 8 active, 1 shared
|
||||||
|
* **Context length:** 128k
|
||||||
|
* **Training Tokens:** 10T
|
||||||
|
* **License:** [OpenMDW-1.1](https://huggingface.co/arcee-ai/Trinity-Mini#license)
|
||||||
|
|
||||||
|
***
|
||||||
|
|
||||||
|
<div align="center">
|
||||||
|
<picture>
|
||||||
|
<img src="https://cdn-uploads.huggingface.co/production/uploads/6435718aaaef013d1aec3b8b/sSVjGNHfrJKmQ6w8I18ek.png" style="background-color:ghostwhite;padding:5px;" width="17%" alt="Powered by Datology">
|
||||||
|
</picture>
|
||||||
|
</div>
|
||||||
|
|
||||||
|
### Running our model
|
||||||
|
|
||||||
|
- [Transformers](https://huggingface.co/arcee-ai/Trinity-Mini#transformers)
|
||||||
|
- [VLLM](https://huggingface.co/arcee-ai/Trinity-Mini#vllm)
|
||||||
|
- [llama.cpp](https://huggingface.co/arcee-ai/Trinity-Mini#llamacpp)
|
||||||
|
- [LM Studio](https://huggingface.co/arcee-ai/Trinity-Mini#lm-studio)
|
||||||
|
|
||||||
|
## Transformers
|
||||||
|
|
||||||
|
Use the `main` transformers branch
|
||||||
|
|
||||||
|
```
|
||||||
|
git clone https://github.com/huggingface/transformers.git
|
||||||
|
cd transformers
|
||||||
|
|
||||||
|
# pip
|
||||||
|
pip install '.[torch]'
|
||||||
|
|
||||||
|
# uv
|
||||||
|
uv pip install '.[torch]'
|
||||||
|
```
|
||||||
|
|
||||||
|
```python
|
||||||
|
from transformers import AutoTokenizer, AutoModelForCausalLM
|
||||||
|
import torch
|
||||||
|
|
||||||
|
model_id = "arcee-ai/Trinity-Nano-Preview"
|
||||||
|
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
||||||
|
model = AutoModelForCausalLM.from_pretrained(
|
||||||
|
model_id,
|
||||||
|
torch_dtype=torch.bfloat16,
|
||||||
|
device_map="auto"
|
||||||
|
)
|
||||||
|
|
||||||
|
messages = [
|
||||||
|
{"role": "user", "content": "Who are you?"},
|
||||||
|
]
|
||||||
|
|
||||||
|
input_ids = tokenizer.apply_chat_template(
|
||||||
|
messages,
|
||||||
|
add_generation_prompt=True,
|
||||||
|
return_tensors="pt"
|
||||||
|
).to(model.device)
|
||||||
|
|
||||||
|
outputs = model.generate(
|
||||||
|
input_ids,
|
||||||
|
max_new_tokens=256,
|
||||||
|
do_sample=True,
|
||||||
|
temperature=0.5,
|
||||||
|
top_k=50,
|
||||||
|
top_p=0.95
|
||||||
|
)
|
||||||
|
|
||||||
|
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
|
||||||
|
print(response)
|
||||||
|
```
|
||||||
|
|
||||||
|
If using a released transformers, simply pass "trust_remote_code=True":
|
||||||
|
|
||||||
|
```python
|
||||||
|
model_id = "arcee-ai/Trinity-Nano-Preview"
|
||||||
|
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
||||||
|
model = AutoModelForCausalLM.from_pretrained(
|
||||||
|
model_id,
|
||||||
|
torch_dtype=torch.bfloat16,
|
||||||
|
device_map="auto",
|
||||||
|
trust_remote_code=True
|
||||||
|
)
|
||||||
|
```
|
||||||
|
|
||||||
|
## VLLM
|
||||||
|
|
||||||
|
Supported in VLLM release 0.11.1
|
||||||
|
|
||||||
|
```
|
||||||
|
# pip
|
||||||
|
pip install "vllm>=0.11.1"
|
||||||
|
```
|
||||||
|
|
||||||
|
Serving the model with suggested settings:
|
||||||
|
|
||||||
|
```
|
||||||
|
vllm serve arcee-train/Trinity-Nano-Preview \
|
||||||
|
--dtype bfloat16 \
|
||||||
|
--enable-auto-tool-choice \
|
||||||
|
--reasoning-parser deepseek_r1 \
|
||||||
|
--tool-call-parser hermes
|
||||||
|
```
|
||||||
|
|
||||||
|
## llama.cpp
|
||||||
|
|
||||||
|
Supported in llama.cpp release b7061
|
||||||
|
|
||||||
|
Download the latest [llama.cpp release](https://github.com/ggml-org/llama.cpp/releases)
|
||||||
|
|
||||||
|
```
|
||||||
|
llama-server -hf arcee-ai/Trinity-Nano-Preview-GGUF:q4_k_m
|
||||||
|
```
|
||||||
|
|
||||||
|
## LM Studio
|
||||||
|
|
||||||
|
Supported in latest LM Studio runtime
|
||||||
|
|
||||||
|
Update to latest available, then verify your runtime by:
|
||||||
|
|
||||||
|
1. Click "Power User" at the bottom left
|
||||||
|
2. Click the green "Developer" icon at the top left
|
||||||
|
3. Select "LM Runtimes" at the top
|
||||||
|
4. Refresh the list of runtimes and verify that the latest is installed
|
||||||
|
|
||||||
|
Then, go to Model Search and search for `arcee-ai/Trinity-Nano-Preview-GGUF`, download your prefered size, and load it up in the chat
|
||||||
|
|
||||||
|
|
||||||
|
## License
|
||||||
|
|
||||||
|
Trinity-Nano-Preview is released under the OpenMDW-1.1 license.
|
||||||
65
chat_template.jinja
Normal file
65
chat_template.jinja
Normal file
@@ -0,0 +1,65 @@
|
|||||||
|
{%- if tools %}
|
||||||
|
{{- '<|im_start|>system\n' }}
|
||||||
|
{%- if messages[0].role == 'system' %}
|
||||||
|
{{- messages[0].content + '\n\n' }}
|
||||||
|
{%- endif %}
|
||||||
|
{{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
||||||
|
{%- for tool in tools %}
|
||||||
|
{{- "\n" }}
|
||||||
|
{{- tool | tojson }}
|
||||||
|
{%- endfor %}
|
||||||
|
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
||||||
|
{%- else %}
|
||||||
|
{%- if messages[0].role == 'system' %}
|
||||||
|
{{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- endif %}
|
||||||
|
{%- for message in messages %}
|
||||||
|
{%- if message.content is string %}
|
||||||
|
{%- set content = message.content %}
|
||||||
|
{%- else %}
|
||||||
|
{%- set content = '' %}
|
||||||
|
{%- endif %}
|
||||||
|
{%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
|
||||||
|
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
||||||
|
{%- elif message.role == "assistant" %}
|
||||||
|
{{- '<|im_start|>' + message.role + '\n' }}
|
||||||
|
{% generation %}
|
||||||
|
{{- content}}
|
||||||
|
{%- if message.tool_calls %}
|
||||||
|
{%- for tool_call in message.tool_calls %}
|
||||||
|
{%- if (loop.first and content) or (not loop.first) %}
|
||||||
|
{{- '\n' }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- if tool_call.function %}
|
||||||
|
{%- set tool_call = tool_call.function %}
|
||||||
|
{%- endif %}
|
||||||
|
{{- '<tool_call>\n{"name": "' }}
|
||||||
|
{{- tool_call.name }}
|
||||||
|
{{- '", "arguments": ' }}
|
||||||
|
{%- if tool_call.arguments is string %}
|
||||||
|
{{- tool_call.arguments }}
|
||||||
|
{%- else %}
|
||||||
|
{{- tool_call.arguments | tojson }}
|
||||||
|
{%- endif %}
|
||||||
|
{{- '}\n</tool_call>' }}
|
||||||
|
{%- endfor %}
|
||||||
|
{%- endif %}
|
||||||
|
{{- '<|im_end|>' }}
|
||||||
|
{% endgeneration%}
|
||||||
|
{{- '\n' }}
|
||||||
|
{%- elif message.role == "tool" %}
|
||||||
|
{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
|
||||||
|
{{- '<|im_start|>user' }}
|
||||||
|
{%- endif %}
|
||||||
|
{{- '\n<tool_response>\n' }}
|
||||||
|
{{- content }}
|
||||||
|
{{- '\n</tool_response>' }}
|
||||||
|
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
||||||
|
{{- '<|im_end|>\n' }}
|
||||||
|
{%- endif %}
|
||||||
|
{%- endif %}
|
||||||
|
{%- endfor %}
|
||||||
|
{%- if add_generation_prompt %}
|
||||||
|
{{- '<|im_start|>assistant\n' }}
|
||||||
|
{%- endif %}
|
||||||
104
config.json
Normal file
104
config.json
Normal file
@@ -0,0 +1,104 @@
|
|||||||
|
{
|
||||||
|
"architectures": [
|
||||||
|
"AfmoeForCausalLM"
|
||||||
|
],
|
||||||
|
"attention_dropout": 0.0,
|
||||||
|
"auto_map": {
|
||||||
|
"AutoConfig": "configuration_afmoe.AfmoeConfig",
|
||||||
|
"AutoModel": "modeling_afmoe.AfmoeModel",
|
||||||
|
"AutoModelForCausalLM": "modeling_afmoe.AfmoeForCausalLM"
|
||||||
|
},
|
||||||
|
"dtype": "bfloat16",
|
||||||
|
"global_attn_every_n_layers": 4,
|
||||||
|
"head_dim": 128,
|
||||||
|
"hidden_act": "silu",
|
||||||
|
"hidden_size": 1024,
|
||||||
|
"initializer_range": 0.02,
|
||||||
|
"intermediate_size": 3072,
|
||||||
|
"layer_types": [
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"full_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"full_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"full_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"full_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"full_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"full_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"full_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"full_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"full_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"full_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"full_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"full_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"full_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"sliding_attention",
|
||||||
|
"full_attention"
|
||||||
|
],
|
||||||
|
"load_balance_coeff": 0.001,
|
||||||
|
"max_position_embeddings": 131072,
|
||||||
|
"model_type": "afmoe",
|
||||||
|
"moe_intermediate_size": 256,
|
||||||
|
"mup_enabled": true,
|
||||||
|
"n_group": 1,
|
||||||
|
"num_attention_heads": 8,
|
||||||
|
"num_dense_layers": 2,
|
||||||
|
"num_expert_groups": 1,
|
||||||
|
"num_experts": 128,
|
||||||
|
"num_experts_per_tok": 8,
|
||||||
|
"num_hidden_layers": 56,
|
||||||
|
"num_key_value_heads": 2,
|
||||||
|
"num_limited_groups": 1,
|
||||||
|
"num_shared_experts": 1,
|
||||||
|
"rms_norm_eps": 1e-05,
|
||||||
|
"rope_scaling": null,
|
||||||
|
"rope_theta": 10000,
|
||||||
|
"route_norm": true,
|
||||||
|
"route_scale": 2.826,
|
||||||
|
"score_func": "sigmoid",
|
||||||
|
"sliding_window": 2048,
|
||||||
|
"tie_word_embeddings": false,
|
||||||
|
"topk_group": 1,
|
||||||
|
"transformers_version": "4.57.3",
|
||||||
|
"use_cache": true,
|
||||||
|
"use_grouped_mm": true,
|
||||||
|
"vocab_size": 200192
|
||||||
|
}
|
||||||
1
configuration.json
Normal file
1
configuration.json
Normal file
@@ -0,0 +1 @@
|
|||||||
|
{"framework": "pytorch", "task": "text-generation", "allow_remote": true}
|
||||||
133
configuration_afmoe.py
Normal file
133
configuration_afmoe.py
Normal file
@@ -0,0 +1,133 @@
|
|||||||
|
# coding=utf-8
|
||||||
|
# Copyright 2022 EleutherAI and the HuggingFace Inc. team. 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.
|
||||||
|
from transformers.configuration_utils import PretrainedConfig
|
||||||
|
from transformers.modeling_rope_utils import rope_config_validation
|
||||||
|
from transformers.configuration_utils import layer_type_validation
|
||||||
|
from transformers.utils import logging
|
||||||
|
|
||||||
|
logger = logging.get_logger(__name__)
|
||||||
|
|
||||||
|
class AfmoeConfig(PretrainedConfig):
|
||||||
|
"""
|
||||||
|
n_group (`int`, *optional*, defaults to 1):
|
||||||
|
Number of groups for routed experts.
|
||||||
|
topk_group (`int`, *optional*, defaults to 1):
|
||||||
|
Number of selected groups for each token(for each token, ensuring the selected experts is only within `topk_group` groups).
|
||||||
|
"""
|
||||||
|
model_type = "afmoe"
|
||||||
|
base_model_pp_plan = {
|
||||||
|
"embed_tokens": (["input_ids"], ["inputs_embeds"]),
|
||||||
|
"layers": (["hidden_states", "attention_mask"], ["hidden_states"]),
|
||||||
|
"norm": (["hidden_states"], ["hidden_states"]),
|
||||||
|
}
|
||||||
|
|
||||||
|
def __init__(
|
||||||
|
self,
|
||||||
|
num_hidden_layers: int = 32,
|
||||||
|
vocab_size: int = 200192,
|
||||||
|
hidden_size: int = 2048,
|
||||||
|
intermediate_size: int = 6144,
|
||||||
|
moe_intermediate_size=1408,
|
||||||
|
num_dense_layers=1,
|
||||||
|
num_attention_heads=16,
|
||||||
|
num_key_value_heads=None,
|
||||||
|
head_dim=128,
|
||||||
|
hidden_act="silu",
|
||||||
|
max_position_embeddings=16384,
|
||||||
|
initializer_range=0.02,
|
||||||
|
rms_norm_eps=1e-5,
|
||||||
|
use_cache=True,
|
||||||
|
tie_word_embeddings=False,
|
||||||
|
rope_theta=10000.0,
|
||||||
|
rope_scaling=None,
|
||||||
|
num_experts=64,
|
||||||
|
num_experts_per_tok=6,
|
||||||
|
num_shared_experts=2,
|
||||||
|
num_expert_groups=1,
|
||||||
|
num_limited_groups=1,
|
||||||
|
score_func="sigmoid",
|
||||||
|
route_norm=True,
|
||||||
|
route_scale=1.0,
|
||||||
|
global_attn_every_n_layers=4,
|
||||||
|
sliding_window=1024,
|
||||||
|
mup_enabled=False,
|
||||||
|
layer_types=None,
|
||||||
|
attention_dropout: float = 0.0,
|
||||||
|
n_group: int = 1,
|
||||||
|
topk_group: int = 1,
|
||||||
|
**kwargs,
|
||||||
|
):
|
||||||
|
self.vocab_size = vocab_size
|
||||||
|
self.max_position_embeddings = max_position_embeddings
|
||||||
|
self.hidden_size = hidden_size
|
||||||
|
self.intermediate_size = intermediate_size
|
||||||
|
self.num_hidden_layers = num_hidden_layers
|
||||||
|
self.num_dense_layers = num_dense_layers
|
||||||
|
self.num_attention_heads = num_attention_heads
|
||||||
|
self.head_dim = head_dim
|
||||||
|
self.hidden_act = hidden_act
|
||||||
|
self.initializer_range = initializer_range
|
||||||
|
self.rms_norm_eps = rms_norm_eps
|
||||||
|
self.use_cache = use_cache
|
||||||
|
self.rope_theta = rope_theta
|
||||||
|
self.rope_scaling = rope_scaling
|
||||||
|
|
||||||
|
|
||||||
|
# MoE specific
|
||||||
|
self.moe_intermediate_size = moe_intermediate_size
|
||||||
|
self.num_experts_per_tok = num_experts_per_tok
|
||||||
|
self.n_group = n_group
|
||||||
|
self.topk_group = topk_group
|
||||||
|
self.num_experts = num_experts
|
||||||
|
self.num_shared_experts = num_shared_experts
|
||||||
|
self.num_expert_groups = num_expert_groups
|
||||||
|
self.num_limited_groups = num_limited_groups
|
||||||
|
self.score_func = score_func
|
||||||
|
self.route_norm = route_norm
|
||||||
|
self.route_scale = route_scale
|
||||||
|
|
||||||
|
|
||||||
|
# Attention specific
|
||||||
|
self.attention_dropout = attention_dropout
|
||||||
|
self.global_attn_every_n_layers = global_attn_every_n_layers
|
||||||
|
self.sliding_window = sliding_window
|
||||||
|
self.layer_types = layer_types
|
||||||
|
if self.layer_types is None:
|
||||||
|
self.layer_types = [
|
||||||
|
"sliding_attention" if bool((i + 1) % global_attn_every_n_layers) else "full_attention" for i in range(self.num_hidden_layers)
|
||||||
|
]
|
||||||
|
layer_type_validation(self.layer_types)
|
||||||
|
|
||||||
|
# muP specific
|
||||||
|
self.mup_enabled = mup_enabled
|
||||||
|
|
||||||
|
if num_key_value_heads is None:
|
||||||
|
num_key_value_heads = num_attention_heads
|
||||||
|
|
||||||
|
self.num_key_value_heads = num_key_value_heads
|
||||||
|
|
||||||
|
|
||||||
|
# Validate rope configs
|
||||||
|
if self.rope_scaling is not None and "type" in self.rope_scaling:
|
||||||
|
self.rope_scaling["rope_type"] = self.rope_scaling["type"]
|
||||||
|
rope_config_validation(self)
|
||||||
|
|
||||||
|
super().__init__(
|
||||||
|
tie_word_embeddings=tie_word_embeddings,
|
||||||
|
**kwargs,
|
||||||
|
)
|
||||||
|
|
||||||
|
|
||||||
|
__all__ = ["AfmoeConfig"]
|
||||||
4
generation_config.json
Normal file
4
generation_config.json
Normal file
@@ -0,0 +1,4 @@
|
|||||||
|
{
|
||||||
|
"_from_model_config": true,
|
||||||
|
"transformers_version": "4.57.3"
|
||||||
|
}
|
||||||
3
model-00001-of-00003.safetensors
Normal file
3
model-00001-of-00003.safetensors
Normal file
@@ -0,0 +1,3 @@
|
|||||||
|
version https://git-lfs.github.com/spec/v1
|
||||||
|
oid sha256:a7d1f9b8c6cfbd5125a1d3a18126f9d7265df41f7024f6ab66ea2aa6146bde8e
|
||||||
|
size 5000898920
|
||||||
3
model-00002-of-00003.safetensors
Normal file
3
model-00002-of-00003.safetensors
Normal file
@@ -0,0 +1,3 @@
|
|||||||
|
version https://git-lfs.github.com/spec/v1
|
||||||
|
oid sha256:a7111078fb9a6a573d59a233e2508b5a07cf3ee52b1300e54925c7fdaa7aa9fe
|
||||||
|
size 5001006952
|
||||||
3
model-00003-of-00003.safetensors
Normal file
3
model-00003-of-00003.safetensors
Normal file
@@ -0,0 +1,3 @@
|
|||||||
|
version https://git-lfs.github.com/spec/v1
|
||||||
|
oid sha256:3f64aad046fc1455143f33b73f1c8e9bf5757a26fb04feee9fb0e83d39a0373a
|
||||||
|
size 2240788480
|
||||||
3
model.safetensors.index.json
Normal file
3
model.safetensors.index.json
Normal file
@@ -0,0 +1,3 @@
|
|||||||
|
version https://git-lfs.github.com/spec/v1
|
||||||
|
oid sha256:43cd417f2e943fefff5b25eba7453f2030d66c6d9c432e27ddac10c13115175f
|
||||||
|
size 1949232
|
||||||
680
modeling_afmoe.py
Normal file
680
modeling_afmoe.py
Normal file
@@ -0,0 +1,680 @@
|
|||||||
|
from typing import Callable, Optional, Tuple, Union
|
||||||
|
|
||||||
|
import torch
|
||||||
|
import torch.nn.functional as F
|
||||||
|
from torch import nn
|
||||||
|
|
||||||
|
from transformers.activations import ACT2FN
|
||||||
|
from transformers.generation import GenerationMixin
|
||||||
|
from transformers.modeling_outputs import (
|
||||||
|
MoeCausalLMOutputWithPast,
|
||||||
|
MoeModelOutputWithPast,
|
||||||
|
)
|
||||||
|
from transformers.modeling_utils import PreTrainedModel, ALL_ATTENTION_FUNCTIONS
|
||||||
|
from transformers.modeling_rope_utils import ROPE_INIT_FUNCTIONS, dynamic_rope_update
|
||||||
|
from transformers.masking_utils import (
|
||||||
|
create_causal_mask,
|
||||||
|
create_sliding_window_causal_mask,
|
||||||
|
)
|
||||||
|
from transformers.modeling_layers import GradientCheckpointingLayer
|
||||||
|
from transformers.processing_utils import Unpack
|
||||||
|
from transformers.utils import TransformersKwargs
|
||||||
|
from transformers.cache_utils import Cache, DynamicCache
|
||||||
|
from transformers.integrations import use_kernel_forward_from_hub
|
||||||
|
|
||||||
|
|
||||||
|
try:
|
||||||
|
from .configuration_afmoe import AfmoeConfig
|
||||||
|
except:
|
||||||
|
from configuration_afmoe import AfmoeConfig
|
||||||
|
|
||||||
|
class AfmoeRotaryEmbedding(nn.Module):
|
||||||
|
|
||||||
|
def __init__(self, config: AfmoeConfig, device=None):
|
||||||
|
super().__init__()
|
||||||
|
# BC: "rope_type" was originally "type"
|
||||||
|
if hasattr(config, "rope_scaling") and config.rope_scaling is not None:
|
||||||
|
self.rope_type = config.rope_scaling.get("rope_type", config.rope_scaling.get("type"))
|
||||||
|
else:
|
||||||
|
self.rope_type = "default"
|
||||||
|
self.max_seq_len_cached = config.max_position_embeddings
|
||||||
|
self.original_max_seq_len = config.max_position_embeddings
|
||||||
|
|
||||||
|
self.config = config
|
||||||
|
self.rope_init_fn = ROPE_INIT_FUNCTIONS[self.rope_type]
|
||||||
|
|
||||||
|
inv_freq, self.attention_scaling = self.rope_init_fn(self.config, device)
|
||||||
|
self.register_buffer("inv_freq", inv_freq, persistent=False)
|
||||||
|
self.original_inv_freq = self.inv_freq
|
||||||
|
|
||||||
|
def _dynamic_frequency_update(self, position_ids, device):
|
||||||
|
"""
|
||||||
|
dynamic RoPE layers should recompute `inv_freq` in the following situations:
|
||||||
|
1 - growing beyond the cached sequence length (allow scaling)
|
||||||
|
2 - the current sequence length is in the original scale (avoid losing precision with small sequences)
|
||||||
|
"""
|
||||||
|
seq_len = torch.max(position_ids) + 1
|
||||||
|
if seq_len > self.max_seq_len_cached: # growth
|
||||||
|
inv_freq, self.attention_scaling = self.rope_init_fn(self.config, device, seq_len=seq_len)
|
||||||
|
self.register_buffer("inv_freq", inv_freq, persistent=False) # TODO joao: may break with compilation
|
||||||
|
self.max_seq_len_cached = seq_len
|
||||||
|
|
||||||
|
if seq_len < self.original_max_seq_len and self.max_seq_len_cached > self.original_max_seq_len: # reset
|
||||||
|
# This .to() is needed if the model has been moved to a device after being initialized (because
|
||||||
|
# the buffer is automatically moved, but not the original copy)
|
||||||
|
self.original_inv_freq = self.original_inv_freq.to(device)
|
||||||
|
self.register_buffer("inv_freq", self.original_inv_freq, persistent=False)
|
||||||
|
self.max_seq_len_cached = self.original_max_seq_len
|
||||||
|
|
||||||
|
@torch.no_grad()
|
||||||
|
def forward(self, x, position_ids):
|
||||||
|
if "dynamic" in self.rope_type:
|
||||||
|
self._dynamic_frequency_update(position_ids, device=x.device)
|
||||||
|
|
||||||
|
# Core RoPE block
|
||||||
|
inv_freq_expanded = self.inv_freq[None, :, None].float().expand(position_ids.shape[0], -1, 1)
|
||||||
|
position_ids_expanded = position_ids[:, None, :].float()
|
||||||
|
# Force float32 (see https://github.com/huggingface/transformers/pull/29285)
|
||||||
|
device_type = x.device.type
|
||||||
|
device_type = device_type if isinstance(device_type, str) and device_type != "mps" else "cpu"
|
||||||
|
with torch.autocast(device_type=device_type, enabled=False):
|
||||||
|
freqs = (inv_freq_expanded.float().to(x.device) @ position_ids_expanded.float()).transpose(1, 2)
|
||||||
|
emb = torch.cat((freqs, freqs), dim=-1)
|
||||||
|
cos = emb.cos()
|
||||||
|
sin = emb.sin()
|
||||||
|
|
||||||
|
# Advanced RoPE types (e.g. yarn) apply a post-processing scaling factor, equivalent to scaling attention
|
||||||
|
cos = cos * self.attention_scaling
|
||||||
|
sin = sin * self.attention_scaling
|
||||||
|
|
||||||
|
return cos.to(dtype=x.dtype), sin.to(dtype=x.dtype)
|
||||||
|
|
||||||
|
|
||||||
|
def rotate_half(x):
|
||||||
|
"""Rotates half the hidden dims of the input."""
|
||||||
|
x1 = x[..., : x.shape[-1] // 2]
|
||||||
|
x2 = x[..., x.shape[-1] // 2 :]
|
||||||
|
return torch.cat((-x2, x1), dim=-1)
|
||||||
|
|
||||||
|
|
||||||
|
def apply_rotary_pos_emb(q, k, cos, sin, position_ids=None, unsqueeze_dim=1):
|
||||||
|
"""Applies Rotary Position Embedding to the query and key tensors.
|
||||||
|
|
||||||
|
Args:
|
||||||
|
q (`torch.Tensor`): The query tensor.
|
||||||
|
k (`torch.Tensor`): The key tensor.
|
||||||
|
cos (`torch.Tensor`): The cosine part of the rotary embedding.
|
||||||
|
sin (`torch.Tensor`): The sine part of the rotary embedding.
|
||||||
|
position_ids (`torch.Tensor`, *optional*):
|
||||||
|
Deprecated and unused.
|
||||||
|
unsqueeze_dim (`int`, *optional*, defaults to 1):
|
||||||
|
The 'unsqueeze_dim' argument specifies the dimension along which to unsqueeze cos[position_ids] and
|
||||||
|
sin[position_ids] so that they can be properly broadcasted to the dimensions of q and k. For example, note
|
||||||
|
that cos[position_ids] and sin[position_ids] have the shape [batch_size, seq_len, head_dim]. Then, if q and
|
||||||
|
k have the shape [batch_size, heads, seq_len, head_dim], then setting unsqueeze_dim=1 makes
|
||||||
|
cos[position_ids] and sin[position_ids] broadcastable to the shapes of q and k. Similarly, if q and k have
|
||||||
|
the shape [batch_size, seq_len, heads, head_dim], then set unsqueeze_dim=2.
|
||||||
|
Returns:
|
||||||
|
`tuple(torch.Tensor)` comprising of the query and key tensors rotated using the Rotary Position Embedding.
|
||||||
|
"""
|
||||||
|
cos = cos.unsqueeze(unsqueeze_dim)
|
||||||
|
sin = sin.unsqueeze(unsqueeze_dim)
|
||||||
|
q_embed = (q * cos) + (rotate_half(q) * sin)
|
||||||
|
k_embed = (k * cos) + (rotate_half(k) * sin)
|
||||||
|
return q_embed, k_embed
|
||||||
|
|
||||||
|
|
||||||
|
def repeat_kv(hidden_states: torch.Tensor, n_rep: int) -> torch.Tensor:
|
||||||
|
"""
|
||||||
|
This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch,
|
||||||
|
num_key_value_heads, seqlen, head_dim) to (batch, num_attention_heads, seqlen, head_dim)
|
||||||
|
"""
|
||||||
|
batch, num_key_value_heads, slen, head_dim = hidden_states.shape
|
||||||
|
if n_rep == 1:
|
||||||
|
return hidden_states
|
||||||
|
hidden_states = hidden_states[:, :, None, :, :].expand(
|
||||||
|
batch, num_key_value_heads, n_rep, slen, head_dim
|
||||||
|
)
|
||||||
|
return hidden_states.reshape(batch, num_key_value_heads * n_rep, slen, head_dim)
|
||||||
|
|
||||||
|
@use_kernel_forward_from_hub("RMSNorm")
|
||||||
|
class AfmoeRMSNorm(nn.Module):
|
||||||
|
def __init__(self, hidden_size: int, eps: float):
|
||||||
|
"""
|
||||||
|
AfmoeRMSNorm is equivalent to T5LayerNorm
|
||||||
|
"""
|
||||||
|
super().__init__()
|
||||||
|
self.weight = nn.Parameter(torch.ones(hidden_size))
|
||||||
|
self.variance_epsilon = eps
|
||||||
|
|
||||||
|
def forward(self, hidden_states):
|
||||||
|
input_dtype = hidden_states.dtype
|
||||||
|
hidden_states = hidden_states.to(torch.float32)
|
||||||
|
variance = hidden_states.pow(2).mean(-1, keepdim=True)
|
||||||
|
hidden_states = hidden_states * torch.rsqrt(variance + self.variance_epsilon)
|
||||||
|
return self.weight * hidden_states.to(input_dtype)
|
||||||
|
|
||||||
|
def extra_repr(self):
|
||||||
|
return f"{tuple(self.weight.shape)}, eps={self.variance_epsilon}"
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
def eager_attention_forward(
|
||||||
|
module: nn.Module,
|
||||||
|
query: torch.Tensor,
|
||||||
|
key: torch.Tensor,
|
||||||
|
value: torch.Tensor,
|
||||||
|
attention_mask: Optional[torch.Tensor],
|
||||||
|
scaling: float,
|
||||||
|
dropout: float = 0.0,
|
||||||
|
**kwargs,
|
||||||
|
):
|
||||||
|
key_states = repeat_kv(key, module.num_key_value_groups)
|
||||||
|
value_states = repeat_kv(value, module.num_key_value_groups)
|
||||||
|
|
||||||
|
attn_weights = torch.matmul(query, key_states.transpose(2, 3)) * scaling
|
||||||
|
if attention_mask is not None:
|
||||||
|
causal_mask = attention_mask[:, :, :, : key_states.shape[-2]]
|
||||||
|
attn_weights = attn_weights + causal_mask
|
||||||
|
|
||||||
|
attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(
|
||||||
|
query.dtype
|
||||||
|
)
|
||||||
|
attn_weights = nn.functional.dropout(
|
||||||
|
attn_weights, p=dropout, training=module.training
|
||||||
|
)
|
||||||
|
attn_output = torch.matmul(attn_weights, value_states)
|
||||||
|
attn_output = attn_output.transpose(1, 2).contiguous()
|
||||||
|
|
||||||
|
return attn_output, attn_weights
|
||||||
|
|
||||||
|
|
||||||
|
class AfmoeMLP(nn.Module):
|
||||||
|
def __init__(self, config, intermediate_size=None):
|
||||||
|
super().__init__()
|
||||||
|
self.config = config
|
||||||
|
self.hidden_size = config.hidden_size
|
||||||
|
self.intermediate_size = intermediate_size or config.intermediate_size
|
||||||
|
self.gate_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
|
||||||
|
self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
|
||||||
|
self.down_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=False)
|
||||||
|
self.act_fn = ACT2FN[config.hidden_act]
|
||||||
|
|
||||||
|
def forward(self, x):
|
||||||
|
return self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x))
|
||||||
|
|
||||||
|
|
||||||
|
class AfmoeTokenChoiceRouter(nn.Module):
|
||||||
|
"""Token-choice top-K router for MoE routing."""
|
||||||
|
|
||||||
|
def __init__(self, config):
|
||||||
|
super().__init__()
|
||||||
|
self.config = config
|
||||||
|
self.top_k = config.num_experts_per_tok
|
||||||
|
self.num_experts = config.num_experts
|
||||||
|
self.score_func = config.score_func
|
||||||
|
self.route_norm = config.route_norm
|
||||||
|
self.route_scale = config.route_scale
|
||||||
|
self.gate = nn.Linear(config.hidden_size, config.num_experts, bias=False)
|
||||||
|
|
||||||
|
def forward(self, hidden_states, expert_bias: torch.Tensor | None):
|
||||||
|
_, _, hidden_dim = hidden_states.shape
|
||||||
|
hidden_states = hidden_states.view(-1, hidden_dim)
|
||||||
|
|
||||||
|
scores = self.gate(hidden_states)
|
||||||
|
|
||||||
|
# Apply scoring function in float32 for stability
|
||||||
|
if self.score_func == "sigmoid":
|
||||||
|
scores = torch.sigmoid(scores.to(torch.float32))
|
||||||
|
else:
|
||||||
|
scores = F.softmax(scores.to(torch.float32), dim=-1)
|
||||||
|
|
||||||
|
if expert_bias is not None:
|
||||||
|
_, selected_experts = torch.topk(scores + expert_bias, k=self.top_k, dim=1)
|
||||||
|
top_scores = scores.gather(dim=1, index=selected_experts)
|
||||||
|
else:
|
||||||
|
top_scores, selected_experts = torch.topk(scores, k=self.top_k, dim=1)
|
||||||
|
|
||||||
|
# Normalize weights if using sigmoid
|
||||||
|
if self.score_func == "sigmoid" and self.route_norm:
|
||||||
|
denominator = top_scores.sum(dim=-1, keepdim=True) + 1e-20
|
||||||
|
top_scores = top_scores / denominator
|
||||||
|
|
||||||
|
top_scores = top_scores * self.route_scale
|
||||||
|
return top_scores, selected_experts
|
||||||
|
|
||||||
|
class AfmoeMoE(nn.Module):
|
||||||
|
def __init__(self, config):
|
||||||
|
super().__init__()
|
||||||
|
self.config = config
|
||||||
|
self.router = AfmoeTokenChoiceRouter(config)
|
||||||
|
|
||||||
|
self.shared_experts = None
|
||||||
|
if config.num_shared_experts > 0:
|
||||||
|
self.shared_experts = AfmoeMLP(
|
||||||
|
config, config.moe_intermediate_size * config.num_shared_experts
|
||||||
|
)
|
||||||
|
self.experts = nn.ModuleList(
|
||||||
|
[AfmoeMLP(
|
||||||
|
config, intermediate_size=config.moe_intermediate_size
|
||||||
|
) for _ in range(config.num_experts)]
|
||||||
|
)
|
||||||
|
self.expert_bias = nn.Parameter(torch.zeros(config.num_experts, dtype=torch.float32), requires_grad=False)
|
||||||
|
|
||||||
|
|
||||||
|
def forward(self, hidden_states):
|
||||||
|
batch_size, seq_len, hidden_dim = hidden_states.shape
|
||||||
|
hidden_states_flat = hidden_states.view(-1, hidden_dim)
|
||||||
|
|
||||||
|
# Get routing decisions
|
||||||
|
top_scores, selected_experts = self.router(hidden_states, self.expert_bias)
|
||||||
|
|
||||||
|
# Process through shared experts
|
||||||
|
if self.shared_experts is not None:
|
||||||
|
shared_output = self.shared_experts(hidden_states_flat)
|
||||||
|
else:
|
||||||
|
shared_output = torch.zeros_like(hidden_states_flat)
|
||||||
|
|
||||||
|
# Reorder tokens by expert for efficient processing
|
||||||
|
token_indices_sorted = torch.argsort(selected_experts.view(-1), stable=True)
|
||||||
|
top_scores_sorted = top_scores.view(-1)[token_indices_sorted]
|
||||||
|
token_to_expert = selected_experts.view(-1)[token_indices_sorted]
|
||||||
|
token_indices_sorted = token_indices_sorted // self.config.num_experts_per_tok
|
||||||
|
|
||||||
|
# Gather input tokens
|
||||||
|
token_indices_expanded = token_indices_sorted.unsqueeze(-1).expand(
|
||||||
|
-1, hidden_dim
|
||||||
|
)
|
||||||
|
routed_input = torch.gather(
|
||||||
|
hidden_states_flat, dim=0, index=token_indices_expanded
|
||||||
|
)
|
||||||
|
|
||||||
|
routed_output = torch.zeros_like(routed_input)
|
||||||
|
for expert_id in range(self.config.num_experts):
|
||||||
|
mask = token_to_expert == expert_id
|
||||||
|
if mask.any():
|
||||||
|
expert_input = routed_input[mask]
|
||||||
|
expert_out = self.experts[expert_id](expert_input)
|
||||||
|
routed_output[mask] = expert_out
|
||||||
|
|
||||||
|
routed_output = (
|
||||||
|
routed_output.to(torch.float32) * top_scores_sorted.unsqueeze(-1)
|
||||||
|
).to(hidden_states.dtype)
|
||||||
|
|
||||||
|
# Scatter back to original positions
|
||||||
|
output = shared_output.scatter_add(
|
||||||
|
dim=0, index=token_indices_expanded, src=routed_output
|
||||||
|
)
|
||||||
|
|
||||||
|
return output.view(batch_size, seq_len, hidden_dim)
|
||||||
|
|
||||||
|
|
||||||
|
class AfmoeAttention(nn.Module):
|
||||||
|
"""Multi-headed attention with local/global pattern and gating."""
|
||||||
|
|
||||||
|
def __init__(self, config: AfmoeConfig, layer_idx: int):
|
||||||
|
super().__init__()
|
||||||
|
self.config = config
|
||||||
|
self.layer_idx = layer_idx
|
||||||
|
self.head_dim = getattr(config, "head_dim", config.hidden_size // config.num_attention_heads)
|
||||||
|
self.num_heads = config.num_attention_heads
|
||||||
|
self.num_key_value_heads = config.num_key_value_heads
|
||||||
|
self.num_key_value_groups = self.num_heads // self.num_key_value_heads
|
||||||
|
|
||||||
|
self.scaling = self.head_dim**-0.5
|
||||||
|
self.attention_dropout = config.attention_dropout
|
||||||
|
self.is_local_attention = config.layer_types[layer_idx] == "sliding_attention"
|
||||||
|
self.sliding_window = config.sliding_window if self.is_local_attention else None
|
||||||
|
|
||||||
|
self.q_proj = nn.Linear(
|
||||||
|
config.hidden_size, self.num_heads * self.head_dim, bias=False
|
||||||
|
)
|
||||||
|
self.k_proj = nn.Linear(
|
||||||
|
config.hidden_size, self.num_key_value_heads * self.head_dim, bias=False
|
||||||
|
)
|
||||||
|
self.v_proj = nn.Linear(
|
||||||
|
config.hidden_size, self.num_key_value_heads * self.head_dim, bias=False
|
||||||
|
)
|
||||||
|
self.o_proj = nn.Linear(
|
||||||
|
self.num_heads * self.head_dim, config.hidden_size, bias=False
|
||||||
|
)
|
||||||
|
|
||||||
|
self.q_norm = AfmoeRMSNorm(self.head_dim, eps=config.rms_norm_eps)
|
||||||
|
self.k_norm = AfmoeRMSNorm(self.head_dim, eps=config.rms_norm_eps)
|
||||||
|
|
||||||
|
self.gate_proj = nn.Linear(
|
||||||
|
config.hidden_size, self.num_heads * self.head_dim, bias=False
|
||||||
|
)
|
||||||
|
|
||||||
|
def forward(
|
||||||
|
self,
|
||||||
|
hidden_states: torch.Tensor,
|
||||||
|
position_embeddings: tuple[torch.Tensor, torch.Tensor],
|
||||||
|
attention_mask: Optional[torch.Tensor],
|
||||||
|
past_key_value: Optional[Cache] = None,
|
||||||
|
cache_position: Optional[torch.LongTensor] = None,
|
||||||
|
**kwargs: Unpack[TransformersKwargs],
|
||||||
|
) -> torch.Tensor:
|
||||||
|
|
||||||
|
input_shape = hidden_states.shape[:-1]
|
||||||
|
hidden_shape = (*input_shape, -1, self.head_dim)
|
||||||
|
|
||||||
|
query_states = self.q_proj(hidden_states).view(hidden_shape)
|
||||||
|
key_states = self.k_proj(hidden_states).view(hidden_shape)
|
||||||
|
value_states = self.v_proj(hidden_states).view(hidden_shape)
|
||||||
|
gate_states = self.gate_proj(hidden_states)
|
||||||
|
|
||||||
|
query_states = self.q_norm(query_states)
|
||||||
|
key_states = self.k_norm(key_states)
|
||||||
|
|
||||||
|
query_states = query_states.transpose(1, 2)
|
||||||
|
key_states = key_states.transpose(1, 2)
|
||||||
|
value_states = value_states.transpose(1, 2)
|
||||||
|
|
||||||
|
if self.is_local_attention:
|
||||||
|
cos, sin = position_embeddings
|
||||||
|
query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin)
|
||||||
|
|
||||||
|
if past_key_value is not None:
|
||||||
|
cache_kwargs = {"cache_position": cache_position}
|
||||||
|
key_states, value_states = past_key_value.update(key_states, value_states, self.layer_idx, cache_kwargs)
|
||||||
|
|
||||||
|
attention_interface: Callable = eager_attention_forward
|
||||||
|
if self.config._attn_implementation != "eager":
|
||||||
|
attention_interface = ALL_ATTENTION_FUNCTIONS[
|
||||||
|
self.config._attn_implementation
|
||||||
|
]
|
||||||
|
|
||||||
|
output, _ = attention_interface(
|
||||||
|
self,
|
||||||
|
query_states,
|
||||||
|
key_states,
|
||||||
|
value_states,
|
||||||
|
attention_mask=attention_mask,
|
||||||
|
dropout=0.0 if not self.training else self.attention_dropout,
|
||||||
|
scaling=self.scaling,
|
||||||
|
sliding_window=self.sliding_window,
|
||||||
|
**kwargs,
|
||||||
|
)
|
||||||
|
|
||||||
|
output = output.view(*input_shape, -1).contiguous()
|
||||||
|
output = output * F.sigmoid(gate_states)
|
||||||
|
return self.o_proj(output)
|
||||||
|
|
||||||
|
|
||||||
|
class AfmoeDecoderLayer(GradientCheckpointingLayer):
|
||||||
|
def __init__(self, config: AfmoeConfig, layer_idx: int):
|
||||||
|
super().__init__()
|
||||||
|
self.hidden_size = config.hidden_size
|
||||||
|
self.layer_idx = layer_idx
|
||||||
|
|
||||||
|
self.self_attn = AfmoeAttention(config=config, layer_idx=layer_idx)
|
||||||
|
self.attention_type = config.layer_types[layer_idx]
|
||||||
|
|
||||||
|
# Dual normalization for attention
|
||||||
|
self.input_layernorm = AfmoeRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
|
||||||
|
self.post_attention_layernorm = AfmoeRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
|
||||||
|
|
||||||
|
# Dual normalization for FFN
|
||||||
|
self.pre_mlp_layernorm = AfmoeRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
|
||||||
|
self.post_mlp_layernorm = AfmoeRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
|
||||||
|
|
||||||
|
# MoE or dense FFN
|
||||||
|
self.moe_enabled = layer_idx >= config.num_dense_layers
|
||||||
|
if self.moe_enabled:
|
||||||
|
self.mlp = AfmoeMoE(config)
|
||||||
|
else:
|
||||||
|
self.mlp = AfmoeMLP(config)
|
||||||
|
|
||||||
|
def forward(
|
||||||
|
self,
|
||||||
|
hidden_states: torch.Tensor,
|
||||||
|
attention_mask: Optional[torch.Tensor] = None,
|
||||||
|
position_ids: Optional[torch.LongTensor] = None,
|
||||||
|
past_key_value: Optional[Cache] = None,
|
||||||
|
use_cache: Optional[bool] = None,
|
||||||
|
cache_position: Optional[torch.LongTensor] = None,
|
||||||
|
position_embeddings: Optional[tuple[torch.Tensor, torch.Tensor]] = None,
|
||||||
|
**kwargs: Unpack[TransformersKwargs],
|
||||||
|
) -> torch.FloatTensor:
|
||||||
|
residual = hidden_states
|
||||||
|
|
||||||
|
# Self Attention with dual normalization
|
||||||
|
hidden_states = self.input_layernorm(hidden_states)
|
||||||
|
hidden_states = self.self_attn(
|
||||||
|
hidden_states=hidden_states,
|
||||||
|
attention_mask=attention_mask,
|
||||||
|
position_ids=position_ids,
|
||||||
|
past_key_value=past_key_value,
|
||||||
|
use_cache=use_cache,
|
||||||
|
cache_position=cache_position,
|
||||||
|
position_embeddings=position_embeddings,
|
||||||
|
**kwargs,
|
||||||
|
)
|
||||||
|
hidden_states = self.post_attention_layernorm(hidden_states)
|
||||||
|
hidden_states = residual + hidden_states
|
||||||
|
|
||||||
|
# FFN with dual normalization
|
||||||
|
residual = hidden_states
|
||||||
|
hidden_states = self.pre_mlp_layernorm(hidden_states)
|
||||||
|
|
||||||
|
if self.moe_enabled:
|
||||||
|
hidden_states = self.mlp(hidden_states)
|
||||||
|
else:
|
||||||
|
hidden_states = self.mlp(hidden_states)
|
||||||
|
|
||||||
|
hidden_states = self.post_mlp_layernorm(hidden_states)
|
||||||
|
hidden_states = residual + hidden_states
|
||||||
|
return hidden_states
|
||||||
|
|
||||||
|
|
||||||
|
class AfmoePreTrainedModel(PreTrainedModel):
|
||||||
|
config_class = AfmoeConfig
|
||||||
|
base_model_prefix = "model"
|
||||||
|
_no_split_modules = ["AfmoeDecoderLayer"]
|
||||||
|
_skip_keys_device_placement = ["past_key_values"]
|
||||||
|
_keep_in_fp32_modules = [
|
||||||
|
"input_layernorm",
|
||||||
|
"post_attention_layernorm",
|
||||||
|
"pre_mlp_layernorm",
|
||||||
|
"post_mlp_layernorm",
|
||||||
|
"q_norm",
|
||||||
|
"k_norm",
|
||||||
|
"norm",
|
||||||
|
]
|
||||||
|
_supports_sdpa = True
|
||||||
|
_supports_attention_backend = True
|
||||||
|
supports_gradient_checkpointing = True
|
||||||
|
|
||||||
|
|
||||||
|
class AfmoeModel(AfmoePreTrainedModel):
|
||||||
|
_no_split_modules = ["AfmoeDecoderLayer"]
|
||||||
|
|
||||||
|
def __init__(self, config: AfmoeConfig):
|
||||||
|
super().__init__(config)
|
||||||
|
self.padding_idx = config.pad_token_id
|
||||||
|
self.vocab_size = config.vocab_size
|
||||||
|
|
||||||
|
self.embed_tokens = nn.Embedding(
|
||||||
|
config.vocab_size, config.hidden_size, self.padding_idx
|
||||||
|
)
|
||||||
|
self.layers = nn.ModuleList(
|
||||||
|
[
|
||||||
|
AfmoeDecoderLayer(config, layer_idx)
|
||||||
|
for layer_idx in range(config.num_hidden_layers)
|
||||||
|
]
|
||||||
|
)
|
||||||
|
self.norm = AfmoeRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
|
||||||
|
self.rotary_emb = AfmoeRotaryEmbedding(config=config)
|
||||||
|
self.gradient_checkpointing = False
|
||||||
|
|
||||||
|
self.post_init()
|
||||||
|
|
||||||
|
def get_input_embeddings(self):
|
||||||
|
return self.embed_tokens
|
||||||
|
|
||||||
|
def set_input_embeddings(self, value):
|
||||||
|
self.embed_tokens = value
|
||||||
|
|
||||||
|
|
||||||
|
def forward(
|
||||||
|
self,
|
||||||
|
input_ids: torch.LongTensor,
|
||||||
|
attention_mask: Optional[torch.Tensor] = None,
|
||||||
|
position_ids: Optional[torch.LongTensor] = None,
|
||||||
|
past_key_values: Optional[list[torch.FloatTensor]] = None,
|
||||||
|
inputs_embeds: Optional[torch.FloatTensor] = None,
|
||||||
|
use_cache: Optional[bool] = None,
|
||||||
|
cache_position: Optional[torch.LongTensor] = None,
|
||||||
|
**kwargs: Unpack[TransformersKwargs],
|
||||||
|
) -> MoeModelOutputWithPast:
|
||||||
|
if (input_ids is None) ^ (inputs_embeds is not None):
|
||||||
|
raise ValueError(
|
||||||
|
"You must specify exactly one of input_ids or inputs_embeds"
|
||||||
|
)
|
||||||
|
|
||||||
|
if use_cache and past_key_values is None:
|
||||||
|
past_key_values = DynamicCache()
|
||||||
|
|
||||||
|
if inputs_embeds is None:
|
||||||
|
inputs_embeds = self.embed_tokens(input_ids)
|
||||||
|
|
||||||
|
if cache_position is None:
|
||||||
|
past_seen_tokens = (
|
||||||
|
past_key_values.get_seq_length() if past_key_values is not None else 0
|
||||||
|
)
|
||||||
|
cache_position = torch.arange(
|
||||||
|
past_seen_tokens,
|
||||||
|
past_seen_tokens + inputs_embeds.shape[1],
|
||||||
|
device=inputs_embeds.device,
|
||||||
|
)
|
||||||
|
if position_ids is None:
|
||||||
|
position_ids = cache_position.unsqueeze(0)
|
||||||
|
|
||||||
|
# It may already have been prepared by e.g. `generate`
|
||||||
|
if not isinstance(causal_mask_mapping := attention_mask, dict):
|
||||||
|
mask_kwargs = {
|
||||||
|
"config": self.config,
|
||||||
|
"input_embeds": inputs_embeds,
|
||||||
|
"attention_mask": attention_mask,
|
||||||
|
"cache_position": cache_position,
|
||||||
|
"past_key_values": past_key_values,
|
||||||
|
}
|
||||||
|
causal_mask_mapping = {
|
||||||
|
"full_attention": create_causal_mask(**mask_kwargs),
|
||||||
|
"sliding_attention": create_sliding_window_causal_mask(**mask_kwargs),
|
||||||
|
}
|
||||||
|
|
||||||
|
hidden_states = inputs_embeds
|
||||||
|
|
||||||
|
# Apply muP input scaling if enabled
|
||||||
|
if self.config.mup_enabled:
|
||||||
|
hidden_states = hidden_states * (self.config.hidden_size**0.5)
|
||||||
|
|
||||||
|
position_embeddings = self.rotary_emb(hidden_states, position_ids)
|
||||||
|
|
||||||
|
for decoder_layer in self.layers:
|
||||||
|
hidden_states = decoder_layer(
|
||||||
|
hidden_states,
|
||||||
|
attention_mask=causal_mask_mapping[decoder_layer.attention_type],
|
||||||
|
position_ids=position_ids,
|
||||||
|
past_key_value=past_key_values,
|
||||||
|
use_cache=use_cache,
|
||||||
|
cache_position=cache_position,
|
||||||
|
position_embeddings=position_embeddings,
|
||||||
|
**kwargs,
|
||||||
|
)
|
||||||
|
|
||||||
|
hidden_states = self.norm(hidden_states)
|
||||||
|
return MoeModelOutputWithPast(
|
||||||
|
last_hidden_state=hidden_states,
|
||||||
|
past_key_values=past_key_values,
|
||||||
|
)
|
||||||
|
|
||||||
|
|
||||||
|
class AfmoeForCausalLM(AfmoePreTrainedModel, GenerationMixin):
|
||||||
|
_tied_weights_keys = ["lm_head.weight"]
|
||||||
|
_tp_plan = {"lm_head": "colwise_rep"}
|
||||||
|
_pp_plan = {"lm_head": (["hidden_states"], ["logits"])}
|
||||||
|
|
||||||
|
def __init__(self, config):
|
||||||
|
super().__init__(config)
|
||||||
|
self.model = AfmoeModel(config)
|
||||||
|
self.vocab_size = config.vocab_size
|
||||||
|
self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
|
||||||
|
|
||||||
|
# Initialize weights and apply final processing
|
||||||
|
self.post_init()
|
||||||
|
|
||||||
|
def get_input_embeddings(self):
|
||||||
|
return self.model.embed_tokens
|
||||||
|
|
||||||
|
def set_input_embeddings(self, value):
|
||||||
|
self.model.embed_tokens = value
|
||||||
|
|
||||||
|
def get_output_embeddings(self):
|
||||||
|
return self.lm_head
|
||||||
|
|
||||||
|
def set_output_embeddings(self, new_embeddings):
|
||||||
|
self.lm_head = new_embeddings
|
||||||
|
|
||||||
|
def set_decoder(self, decoder):
|
||||||
|
self.model = decoder
|
||||||
|
|
||||||
|
def get_decoder(self):
|
||||||
|
return self.model
|
||||||
|
|
||||||
|
def forward(
|
||||||
|
self,
|
||||||
|
input_ids: torch.LongTensor,
|
||||||
|
attention_mask: Optional[torch.Tensor] = None,
|
||||||
|
position_ids: Optional[torch.LongTensor] = None,
|
||||||
|
past_key_values: Optional[Cache] = None,
|
||||||
|
inputs_embeds: Optional[torch.FloatTensor] = None,
|
||||||
|
labels: Optional[torch.LongTensor] = None,
|
||||||
|
use_cache: Optional[bool] = None,
|
||||||
|
cache_position: Optional[torch.LongTensor] = None,
|
||||||
|
logits_to_keep: Union[int, torch.Tensor] = 0,
|
||||||
|
token_type_ids: Optional[torch.Tensor] = None, # will be ignored
|
||||||
|
**kwargs: Unpack[TransformersKwargs],
|
||||||
|
) -> Union[Tuple, MoeCausalLMOutputWithPast]:
|
||||||
|
outputs: MoeModelOutputWithPast = self.model(
|
||||||
|
input_ids=input_ids,
|
||||||
|
attention_mask=attention_mask,
|
||||||
|
position_ids=position_ids,
|
||||||
|
past_key_values=past_key_values,
|
||||||
|
inputs_embeds=inputs_embeds,
|
||||||
|
use_cache=use_cache,
|
||||||
|
cache_position=cache_position,
|
||||||
|
**kwargs,
|
||||||
|
)
|
||||||
|
|
||||||
|
hidden_states = outputs.last_hidden_state
|
||||||
|
# Only compute necessary logits
|
||||||
|
slice_indices = (
|
||||||
|
slice(-logits_to_keep, None)
|
||||||
|
if isinstance(logits_to_keep, int)
|
||||||
|
else logits_to_keep
|
||||||
|
)
|
||||||
|
logits = self.lm_head(hidden_states[:, slice_indices, :])
|
||||||
|
|
||||||
|
loss = None
|
||||||
|
if labels is not None:
|
||||||
|
loss = self.loss_function(logits, labels, self.vocab_size, **kwargs)
|
||||||
|
|
||||||
|
|
||||||
|
return MoeCausalLMOutputWithPast(
|
||||||
|
loss=loss,
|
||||||
|
logits=logits,
|
||||||
|
past_key_values=outputs.past_key_values,
|
||||||
|
hidden_states=outputs.hidden_states,
|
||||||
|
attentions=outputs.attentions,
|
||||||
|
router_logits=outputs.router_logits,
|
||||||
|
)
|
||||||
|
|
||||||
|
|
||||||
|
__all__ = [
|
||||||
|
"AfmoeForCausalLM",
|
||||||
|
"AfmoeModel",
|
||||||
|
"AfmoePreTrainedModel",
|
||||||
|
]
|
||||||
23
special_tokens_map.json
Normal file
23
special_tokens_map.json
Normal file
@@ -0,0 +1,23 @@
|
|||||||
|
{
|
||||||
|
"bos_token": {
|
||||||
|
"content": "<|begin_of_text|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false
|
||||||
|
},
|
||||||
|
"eos_token": {
|
||||||
|
"content": "<|im_end|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false
|
||||||
|
},
|
||||||
|
"pad_token": {
|
||||||
|
"content": "<|pad|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false
|
||||||
|
}
|
||||||
|
}
|
||||||
3
tokenizer.json
Normal file
3
tokenizer.json
Normal file
@@ -0,0 +1,3 @@
|
|||||||
|
version https://git-lfs.github.com/spec/v1
|
||||||
|
oid sha256:606b0740c5d8940dbee3268e05beaa3e34f4e43bc7df153681122f8fcc78e134
|
||||||
|
size 14614897
|
||||||
271
tokenizer_config.json
Normal file
271
tokenizer_config.json
Normal file
@@ -0,0 +1,271 @@
|
|||||||
|
{
|
||||||
|
"add_bos_token": false,
|
||||||
|
"add_eos_token": false,
|
||||||
|
"add_prefix_space": null,
|
||||||
|
"added_tokens_decoder": {
|
||||||
|
"0": {
|
||||||
|
"content": "<|begin_of_text|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"1": {
|
||||||
|
"content": "<|end_of_text|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"2": {
|
||||||
|
"content": "<|im_start|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"3": {
|
||||||
|
"content": "<|im_end|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"4": {
|
||||||
|
"content": "<|eot_id|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"5": {
|
||||||
|
"content": "<|start|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"6": {
|
||||||
|
"content": "<|channel|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"7": {
|
||||||
|
"content": "<|message|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"8": {
|
||||||
|
"content": "<|end|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"9": {
|
||||||
|
"content": "<|fitm_start|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"10": {
|
||||||
|
"content": "<|fitm_end|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"11": {
|
||||||
|
"content": "<|fitm_hole|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"12": {
|
||||||
|
"content": "<|pad|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"13": {
|
||||||
|
"content": "<|reserved_special_0|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"14": {
|
||||||
|
"content": "<|reserved_special_1|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"15": {
|
||||||
|
"content": "<|reserved_special_2|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"16": {
|
||||||
|
"content": "<|reserved_special_3|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"17": {
|
||||||
|
"content": "<|reserved_special_4|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"18": {
|
||||||
|
"content": "<|reserved_special_5|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"19": {
|
||||||
|
"content": "<|reserved_special_6|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"20": {
|
||||||
|
"content": "<|reserved_special_7|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"21": {
|
||||||
|
"content": "<|reserved_special_8|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"22": {
|
||||||
|
"content": "<|reserved_special_9|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"23": {
|
||||||
|
"content": "<|reserved_special_10|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"24": {
|
||||||
|
"content": "<|reserved_special_11|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"25": {
|
||||||
|
"content": "<|reserved_special_12|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"26": {
|
||||||
|
"content": "<|reserved_special_13|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"27": {
|
||||||
|
"content": "<|reserved_special_14|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"28": {
|
||||||
|
"content": "<|reserved_special_15|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"29": {
|
||||||
|
"content": "<|reserved_special_16|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"30": {
|
||||||
|
"content": "<|reserved_special_17|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"31": {
|
||||||
|
"content": "<|reserved_special_18|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"bos_token": "<|begin_of_text|>",
|
||||||
|
"clean_up_tokenization_spaces": false,
|
||||||
|
"eos_token": "<|im_end|>",
|
||||||
|
"extra_special_tokens": {},
|
||||||
|
"model_max_length": 65536,
|
||||||
|
"pad_token": "<|pad|>",
|
||||||
|
"tokenizer_class": "PreTrainedTokenizerFast",
|
||||||
|
"use_default_system_prompt": false
|
||||||
|
}
|
||||||
Reference in New Issue
Block a user