Commit
•
39c1184
1
Parent(s):
2588440
New weights for the model
Browse files- README.md +0 -132
- config.json +71 -73
- configuration_phi3.py +227 -213
- model-00001-of-00002.safetensors +1 -1
- model-00002-of-00002.safetensors +1 -1
- modeling_phi3.py +0 -0
- tokenizer.json +1 -29
- tokenizer_config.json +4 -2
README.md
DELETED
@@ -1,132 +0,0 @@
|
|
1 |
-
---
|
2 |
-
license: apache-2.0
|
3 |
-
language:
|
4 |
-
- en
|
5 |
-
library_name: transformers
|
6 |
-
pipeline_tag: text2text-generation
|
7 |
-
tags:
|
8 |
-
- code
|
9 |
-
- sql
|
10 |
-
- text-to-sql
|
11 |
-
---
|
12 |
-
|
13 |
-
Introducing Hrida-T2SQL-3B-128k-V0.1, our latest small language model (SLM) tailored for data scientists and industry professionals. This advanced model marks a significant upgrade from our previous release, now equipped with an expanded 128k token context window for handling even the most intricate data queries with precision. Powered by the Phi 3 architecture, it effortlessly converts natural language queries into precise SQL commands, enhancing data analysis efficiency and decision-making capabilities.
|
14 |
-
|
15 |
-
For full details of this model please read our [blog post](https://www.hridaai.com/blog/t2sql).
|
16 |
-
|
17 |
-
|
18 |
-
## Prompt Template
|
19 |
-
|
20 |
-
```txt
|
21 |
-
### Instruction:
|
22 |
-
Provide the system prompt.
|
23 |
-
|
24 |
-
### Dialect:
|
25 |
-
Specify the SQL dialect (e.g., MySQL, PostgreSQL, SQL Server, etc.).
|
26 |
-
|
27 |
-
### Context:
|
28 |
-
Provide the database schema including table names, column names, and data types.
|
29 |
-
|
30 |
-
### Input:
|
31 |
-
User's query.
|
32 |
-
|
33 |
-
### Response:
|
34 |
-
Expected SQL query output based on the input and context.
|
35 |
-
|
36 |
-
```
|
37 |
-
|
38 |
-
- **Instruction (System Prompt)**: This guides the model on processing input to generate the SQL query response effectively.
|
39 |
-
- **Dialect (Optional)**: Specify the SQL variant the model should use to ensure the generated query conforms to the correct syntax.
|
40 |
-
- **Context**: Provide the database schema to the model for generating accurate SQL queries.
|
41 |
-
- **Input**: Provide the user query for the model to comprehend and transform into an SQL query.
|
42 |
-
- **Response**: Expected output from the model.
|
43 |
-
|
44 |
-
|
45 |
-
## Chat Prompt Template
|
46 |
-
|
47 |
-
```txt
|
48 |
-
<s>
|
49 |
-
<|system|>
|
50 |
-
{ Instruction / System Prompt }
|
51 |
-
<|user|>
|
52 |
-
{ Context / User Query } <|end|>
|
53 |
-
<|assistant|>
|
54 |
-
```
|
55 |
-
|
56 |
-
## Run the Model
|
57 |
-
|
58 |
-
### Using Transformers
|
59 |
-
|
60 |
-
```python
|
61 |
-
import torch
|
62 |
-
from transformers import AutoModelForCausalLM, AutoTokenizer
|
63 |
-
|
64 |
-
# Define the model and tokenizer
|
65 |
-
model_id = "HridaAI/Hrida-T2SQL-3B-128k-V0.1"
|
66 |
-
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
|
67 |
-
model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.float16, trust_remote_code=True)
|
68 |
-
|
69 |
-
# Define the context and prompt
|
70 |
-
prompt = """
|
71 |
-
Answer to the query will be in the form of an SQL query.
|
72 |
-
### Context: CREATE TABLE Employees (
|
73 |
-
EmployeeID INT PRIMARY KEY,
|
74 |
-
FirstName VARCHAR(50),
|
75 |
-
LastName VARCHAR(50),
|
76 |
-
Age INT,
|
77 |
-
DepartmentID INT,
|
78 |
-
Salary DECIMAL(10, 2),
|
79 |
-
DateHired DATE,
|
80 |
-
Active BOOLEAN,
|
81 |
-
FOREIGN KEY (DepartmentID) REFERENCES Departments(DepartmentID)
|
82 |
-
);
|
83 |
-
|
84 |
-
CREATE TABLE Departments (
|
85 |
-
DepartmentID INT PRIMARY KEY,
|
86 |
-
DepartmentName VARCHAR(100),
|
87 |
-
Location VARCHAR(100)
|
88 |
-
);
|
89 |
-
### Input: Write a SQL query to select all the employees who are active.
|
90 |
-
### Response:
|
91 |
-
"""
|
92 |
-
# Prepare the input
|
93 |
-
messages = [{"role": "user", "content": prompt}]
|
94 |
-
inputs = tokenizer.apply_chat_template(messages, return_tensors="pt", add_generation_prompt=True)
|
95 |
-
|
96 |
-
# Generate the output
|
97 |
-
outputs = model.generate(inputs, max_length=300)
|
98 |
-
print(tokenizer.decode(outputs[0]))
|
99 |
-
|
100 |
-
|
101 |
-
```
|
102 |
-
|
103 |
-
### Using MLX
|
104 |
-
|
105 |
-
```python
|
106 |
-
from mlx_lm import generate, load
|
107 |
-
|
108 |
-
model,tokenizer = load("HridaAI/Hrida-T2SQL-3B-128k-V0.1")
|
109 |
-
|
110 |
-
prompt = """
|
111 |
-
Answer to the quey will be in the form of SQL query.
|
112 |
-
### Context: CREATE TABLE Employees (
|
113 |
-
EmployeeID INT PRIMARY KEY,
|
114 |
-
FirstName VARCHAR(50),
|
115 |
-
LastName VARCHAR(50),
|
116 |
-
Age INT,
|
117 |
-
DepartmentID INT,
|
118 |
-
Salary DECIMAL(10, 2),
|
119 |
-
DateHired DATE,
|
120 |
-
Active BOOLEAN,
|
121 |
-
FOREIGN KEY (DepartmentID) REFERENCES Departments(DepartmentID)
|
122 |
-
);
|
123 |
-
|
124 |
-
CREATE TABLE Departments (
|
125 |
-
DepartmentID INT PRIMARY KEY,
|
126 |
-
DepartmentName VARCHAR(100),
|
127 |
-
Location VARCHAR(100)
|
128 |
-
); ### Input: Write a SQL query to select all the employees who are active. ### Response:"""
|
129 |
-
|
130 |
-
response = generate(model=model,tokenizer=tokenizer,prompt=prompt, verbose=True)
|
131 |
-
|
132 |
-
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
config.json
CHANGED
@@ -6,9 +6,7 @@
|
|
6 |
"attention_dropout": 0.0,
|
7 |
"auto_map": {
|
8 |
"AutoConfig": "configuration_phi3.Phi3Config",
|
9 |
-
"AutoModelForCausalLM": "modeling_phi3.Phi3ForCausalLM"
|
10 |
-
"AutoModelForSequenceClassification": "modeling_phi3.Phi3ForSequenceClassification",
|
11 |
-
"AutoModelForTokenClassification": "modeling_phi3.Phi3ForTokenClassification"
|
12 |
},
|
13 |
"bos_token_id": 1,
|
14 |
"embd_pdrop": 0.0,
|
@@ -28,71 +26,64 @@
|
|
28 |
"rms_norm_eps": 1e-05,
|
29 |
"rope_scaling": {
|
30 |
"long_factor": [
|
31 |
-
1.
|
32 |
-
1.
|
33 |
-
1.
|
34 |
-
1.
|
35 |
-
1.
|
36 |
-
1.
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
-
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
-
|
63 |
-
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
63.
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
-
64.
|
75 |
-
64.
|
76 |
-
64.
|
77 |
-
64.
|
78 |
-
64.
|
79 |
],
|
80 |
"short_factor": [
|
81 |
-
1.05,
|
82 |
-
1.05,
|
83 |
-
1.05,
|
84 |
1.1,
|
85 |
1.1,
|
86 |
-
1.
|
87 |
-
1.2000000000000002,
|
88 |
-
1.2500000000000002,
|
89 |
1.3000000000000003,
|
90 |
1.3500000000000003,
|
91 |
-
1.
|
92 |
-
|
93 |
-
|
94 |
-
2.000000000000001,
|
95 |
-
2.000000000000001,
|
96 |
2.000000000000001,
|
97 |
2.000000000000001,
|
98 |
2.000000000000001,
|
@@ -113,27 +104,34 @@
|
|
113 |
2.0500000000000007,
|
114 |
2.0500000000000007,
|
115 |
2.0500000000000007,
|
|
|
|
|
|
|
116 |
2.1000000000000005,
|
117 |
2.1000000000000005,
|
118 |
-
2.1000000000000005,
|
119 |
-
2.1500000000000004,
|
120 |
2.1500000000000004,
|
121 |
-
2.
|
122 |
-
2.
|
123 |
-
2.
|
124 |
-
2.
|
|
|
|
|
|
|
|
|
|
|
|
|
125 |
2.7499999999999982,
|
126 |
-
2.
|
127 |
-
2.
|
128 |
-
|
129 |
],
|
130 |
-
"type": "
|
131 |
},
|
132 |
"rope_theta": 10000.0,
|
133 |
"sliding_window": 262144,
|
134 |
"tie_word_embeddings": false,
|
135 |
"torch_dtype": "bfloat16",
|
136 |
-
"transformers_version": "4.
|
137 |
"use_cache": true,
|
138 |
"vocab_size": 32064
|
139 |
}
|
|
|
6 |
"attention_dropout": 0.0,
|
7 |
"auto_map": {
|
8 |
"AutoConfig": "configuration_phi3.Phi3Config",
|
9 |
+
"AutoModelForCausalLM": "modeling_phi3.Phi3ForCausalLM"
|
|
|
|
|
10 |
},
|
11 |
"bos_token_id": 1,
|
12 |
"embd_pdrop": 0.0,
|
|
|
26 |
"rms_norm_eps": 1e-05,
|
27 |
"rope_scaling": {
|
28 |
"long_factor": [
|
29 |
+
1.0700000524520874,
|
30 |
+
1.1200000047683716,
|
31 |
+
1.149999976158142,
|
32 |
+
1.4199999570846558,
|
33 |
+
1.5699999332427979,
|
34 |
+
1.7999999523162842,
|
35 |
+
2.129999876022339,
|
36 |
+
2.129999876022339,
|
37 |
+
3.009999990463257,
|
38 |
+
5.910000324249268,
|
39 |
+
6.950000286102295,
|
40 |
+
9.070000648498535,
|
41 |
+
9.930000305175781,
|
42 |
+
10.710000038146973,
|
43 |
+
11.130000114440918,
|
44 |
+
14.609999656677246,
|
45 |
+
15.409998893737793,
|
46 |
+
19.809999465942383,
|
47 |
+
37.279998779296875,
|
48 |
+
38.279998779296875,
|
49 |
+
38.599998474121094,
|
50 |
+
40.12000274658203,
|
51 |
+
46.20000457763672,
|
52 |
+
50.940006256103516,
|
53 |
+
53.66000747680664,
|
54 |
+
54.9373893737793,
|
55 |
+
56.89738845825195,
|
56 |
+
57.28738784790039,
|
57 |
+
59.98738479614258,
|
58 |
+
60.86738586425781,
|
59 |
+
60.887386322021484,
|
60 |
+
61.71739196777344,
|
61 |
+
62.91739273071289,
|
62 |
+
62.957393646240234,
|
63 |
+
63.41739273071289,
|
64 |
+
63.8173942565918,
|
65 |
+
63.83739471435547,
|
66 |
+
63.897396087646484,
|
67 |
+
63.93739700317383,
|
68 |
+
64.06739807128906,
|
69 |
+
64.11434936523438,
|
70 |
+
64.12435150146484,
|
71 |
+
64.15435028076172,
|
72 |
+
64.19435119628906,
|
73 |
+
64.24435424804688,
|
74 |
+
64.57435607910156,
|
75 |
+
64.69000244140625,
|
76 |
+
64.76000213623047
|
77 |
],
|
78 |
"short_factor": [
|
|
|
|
|
|
|
79 |
1.1,
|
80 |
1.1,
|
81 |
+
1.1,
|
|
|
|
|
82 |
1.3000000000000003,
|
83 |
1.3500000000000003,
|
84 |
+
1.3500000000000003,
|
85 |
+
1.4000000000000004,
|
86 |
+
1.5500000000000005,
|
|
|
|
|
87 |
2.000000000000001,
|
88 |
2.000000000000001,
|
89 |
2.000000000000001,
|
|
|
104 |
2.0500000000000007,
|
105 |
2.0500000000000007,
|
106 |
2.0500000000000007,
|
107 |
+
2.0500000000000007,
|
108 |
+
2.0500000000000007,
|
109 |
+
2.0500000000000007,
|
110 |
2.1000000000000005,
|
111 |
2.1000000000000005,
|
|
|
|
|
112 |
2.1500000000000004,
|
113 |
+
2.25,
|
114 |
+
2.25,
|
115 |
+
2.25,
|
116 |
+
2.25,
|
117 |
+
2.25,
|
118 |
+
2.3999999999999995,
|
119 |
+
2.4499999999999993,
|
120 |
+
2.499999999999999,
|
121 |
+
2.6999999999999984,
|
122 |
+
2.6999999999999984,
|
123 |
2.7499999999999982,
|
124 |
+
2.799999999999998,
|
125 |
+
2.8999999999999977,
|
126 |
+
3.049999999999997
|
127 |
],
|
128 |
+
"type": "longrope"
|
129 |
},
|
130 |
"rope_theta": 10000.0,
|
131 |
"sliding_window": 262144,
|
132 |
"tie_word_embeddings": false,
|
133 |
"torch_dtype": "bfloat16",
|
134 |
+
"transformers_version": "4.40.2",
|
135 |
"use_cache": true,
|
136 |
"vocab_size": 32064
|
137 |
}
|
configuration_phi3.py
CHANGED
@@ -1,213 +1,227 @@
|
|
1 |
-
# coding=utf-8
|
2 |
-
# Copyright 2024 Microsoft and the HuggingFace Inc. team. All rights reserved.
|
3 |
-
#
|
4 |
-
# Licensed under the Apache License, Version 2.0 (the "License");
|
5 |
-
# you may not use this file except in compliance with the License.
|
6 |
-
# You may obtain a copy of the License at
|
7 |
-
#
|
8 |
-
# http://www.apache.org/licenses/LICENSE-2.0
|
9 |
-
#
|
10 |
-
# Unless required by applicable law or agreed to in writing, software
|
11 |
-
# distributed under the License is distributed on an "AS IS" BASIS,
|
12 |
-
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13 |
-
# See the License for the specific language governing permissions and
|
14 |
-
# limitations under the License.
|
15 |
-
|
16 |
-
""" Phi-3 model configuration"""
|
17 |
-
|
18 |
-
|
19 |
-
from transformers.configuration_utils import PretrainedConfig
|
20 |
-
from transformers.utils import logging
|
21 |
-
|
22 |
-
|
23 |
-
logger = logging.get_logger(__name__)
|
24 |
-
|
25 |
-
PHI3_PRETRAINED_CONFIG_ARCHIVE_MAP = {
|
26 |
-
"microsoft/Phi-3-mini-4k-instruct": "https://huggingface.co/microsoft/Phi-3-mini-4k-instruct/resolve/main/config.json",
|
27 |
-
"microsoft/Phi-3-mini-128k-instruct": "https://huggingface.co/microsoft/Phi-3-mini-128k-instruct/resolve/main/config.json",
|
28 |
-
}
|
29 |
-
|
30 |
-
|
31 |
-
class Phi3Config(PretrainedConfig):
|
32 |
-
r"""
|
33 |
-
This is the configuration class to store the configuration of a [`Phi3Model`]. It is used to instantiate a Phi-3
|
34 |
-
model according to the specified arguments, defining the model architecture. Instantiating a configuration with the
|
35 |
-
defaults will yield a similar configuration to that of the
|
36 |
-
[microsoft/Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct).
|
37 |
-
|
38 |
-
Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
|
39 |
-
documentation from [`PretrainedConfig`] for more information.
|
40 |
-
|
41 |
-
Args:
|
42 |
-
vocab_size (`int`, *optional*, defaults to 32064):
|
43 |
-
Vocabulary size of the Phi-3 model. Defines the number of different tokens that can be represented by the
|
44 |
-
`inputs_ids` passed when calling [`Phi3Model`].
|
45 |
-
hidden_size (`int`, *optional*, defaults to 3072):
|
46 |
-
Dimension of the hidden representations.
|
47 |
-
intermediate_size (`int`, *optional*, defaults to 8192):
|
48 |
-
Dimension of the MLP representations.
|
49 |
-
num_hidden_layers (`int`, *optional*, defaults to 32):
|
50 |
-
Number of hidden layers in the Transformer decoder.
|
51 |
-
num_attention_heads (`int`, *optional*, defaults to 32):
|
52 |
-
Number of attention heads for each attention layer in the Transformer decoder.
|
53 |
-
num_key_value_heads (`int`, *optional*):
|
54 |
-
This is the number of key_value heads that should be used to implement Grouped Query Attention. If
|
55 |
-
`num_key_value_heads=num_attention_heads`, the model will use Multi Head Attention (MHA), if
|
56 |
-
`num_key_value_heads=1 the model will use Multi Query Attention (MQA) otherwise GQA is used. When
|
57 |
-
converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed
|
58 |
-
by meanpooling all the original heads within that group. For more details checkout [this
|
59 |
-
paper](https://arxiv.org/pdf/2305.13245.pdf). If it is not specified, will default to
|
60 |
-
`num_attention_heads`.
|
61 |
-
resid_pdrop (`float`, *optional*, defaults to 0.0):
|
62 |
-
Dropout probability for mlp outputs.
|
63 |
-
embd_pdrop (`int`, *optional*, defaults to 0.0):
|
64 |
-
The dropout ratio for the embeddings.
|
65 |
-
attention_dropout (`float`, *optional*, defaults to 0.0):
|
66 |
-
The dropout ratio after computing the attention scores.
|
67 |
-
hidden_act (`str` or `function`, *optional*, defaults to `"silu"`):
|
68 |
-
The non-linear activation function (function or string) in the decoder.
|
69 |
-
max_position_embeddings (`int`, *optional*, defaults to 4096):
|
70 |
-
The maximum sequence length that this model might ever be used with.
|
71 |
-
original_max_position_embeddings (`int`, *optional*, defaults to 4096):
|
72 |
-
The maximum sequence length that this model was trained with. This is used to determine the size of the
|
73 |
-
original RoPE embeddings when using long scaling.
|
74 |
-
initializer_range (`float`, *optional*, defaults to 0.02):
|
75 |
-
The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
|
76 |
-
rms_norm_eps (`float`, *optional*, defaults to 1e-05):
|
77 |
-
The epsilon value used for the RMSNorm.
|
78 |
-
use_cache (`bool`, *optional*, defaults to `True`):
|
79 |
-
Whether or not the model should return the last key/values attentions (not used by all models). Only
|
80 |
-
relevant if `config.is_decoder=True`. Whether to tie weight embeddings or not.
|
81 |
-
tie_word_embeddings (`bool`, *optional*, defaults to `False`):
|
82 |
-
Whether to tie weight embeddings
|
83 |
-
rope_theta (`float`, *optional*, defaults to 10000.0):
|
84 |
-
The base period of the RoPE embeddings.
|
85 |
-
rope_scaling (`dict`, *optional*):
|
86 |
-
The scaling strategy for the RoPE embeddings. If `None`, no scaling is applied. If a dictionary, it must
|
87 |
-
contain the following keys: `type`, `short_factor` and `long_factor`. The `type` must be
|
88 |
-
the `short_factor` and `long_factor` must be lists of numbers with the same length as the hidden size
|
89 |
-
divided by the number of attention heads divided by 2.
|
90 |
-
bos_token_id (`int`, *optional*, defaults to 1):
|
91 |
-
The id of the "beginning-of-sequence" token.
|
92 |
-
eos_token_id (`int`, *optional*, defaults to 32000):
|
93 |
-
The id of the "end-of-sequence" token.
|
94 |
-
pad_token_id (`int`, *optional*, defaults to 32000):
|
95 |
-
The id of the padding token.
|
96 |
-
sliding_window (`int`, *optional*):
|
97 |
-
Sliding window attention window size. If `None`, no sliding window is applied.
|
98 |
-
|
99 |
-
Example:
|
100 |
-
|
101 |
-
```python
|
102 |
-
>>> from transformers import Phi3Model, Phi3Config
|
103 |
-
|
104 |
-
>>> # Initializing a Phi-3 style configuration
|
105 |
-
>>> configuration = Phi3Config.from_pretrained("microsoft/Phi-3-mini-4k-instruct")
|
106 |
-
|
107 |
-
>>> # Initializing a model from the configuration
|
108 |
-
>>> model = Phi3Model(configuration)
|
109 |
-
|
110 |
-
>>> # Accessing the model configuration
|
111 |
-
>>> configuration = model.config
|
112 |
-
```"""
|
113 |
-
|
114 |
-
model_type = "phi3"
|
115 |
-
keys_to_ignore_at_inference = ["past_key_values"]
|
116 |
-
|
117 |
-
def __init__(
|
118 |
-
self,
|
119 |
-
vocab_size=32064,
|
120 |
-
hidden_size=3072,
|
121 |
-
intermediate_size=8192,
|
122 |
-
num_hidden_layers=32,
|
123 |
-
num_attention_heads=32,
|
124 |
-
num_key_value_heads=None,
|
125 |
-
resid_pdrop=0.0,
|
126 |
-
embd_pdrop=0.0,
|
127 |
-
attention_dropout=0.0,
|
128 |
-
hidden_act="silu",
|
129 |
-
max_position_embeddings=4096,
|
130 |
-
original_max_position_embeddings=4096,
|
131 |
-
initializer_range=0.02,
|
132 |
-
rms_norm_eps=1e-5,
|
133 |
-
use_cache=True,
|
134 |
-
tie_word_embeddings=False,
|
135 |
-
rope_theta=10000.0,
|
136 |
-
rope_scaling=None,
|
137 |
-
bos_token_id=1,
|
138 |
-
eos_token_id=32000,
|
139 |
-
pad_token_id=32000,
|
140 |
-
sliding_window=None,
|
141 |
-
**kwargs,
|
142 |
-
):
|
143 |
-
self.vocab_size = vocab_size
|
144 |
-
self.hidden_size = hidden_size
|
145 |
-
self.intermediate_size = intermediate_size
|
146 |
-
self.num_hidden_layers = num_hidden_layers
|
147 |
-
self.num_attention_heads = num_attention_heads
|
148 |
-
|
149 |
-
if num_key_value_heads is None:
|
150 |
-
num_key_value_heads = num_attention_heads
|
151 |
-
|
152 |
-
self.num_key_value_heads = num_key_value_heads
|
153 |
-
self.resid_pdrop = resid_pdrop
|
154 |
-
self.embd_pdrop = embd_pdrop
|
155 |
-
self.attention_dropout = attention_dropout
|
156 |
-
self.hidden_act = hidden_act
|
157 |
-
self.max_position_embeddings = max_position_embeddings
|
158 |
-
self.original_max_position_embeddings = original_max_position_embeddings
|
159 |
-
self.initializer_range = initializer_range
|
160 |
-
self.rms_norm_eps = rms_norm_eps
|
161 |
-
self.use_cache = use_cache
|
162 |
-
self.rope_theta = rope_theta
|
163 |
-
self.rope_scaling = rope_scaling
|
164 |
-
self.
|
165 |
-
self.
|
166 |
-
|
167 |
-
|
168 |
-
|
169 |
-
|
170 |
-
|
171 |
-
|
172 |
-
|
173 |
-
|
174 |
-
|
175 |
-
|
176 |
-
|
177 |
-
|
178 |
-
|
179 |
-
|
180 |
-
|
181 |
-
|
182 |
-
|
183 |
-
|
184 |
-
|
185 |
-
|
186 |
-
|
187 |
-
|
188 |
-
|
189 |
-
|
190 |
-
|
191 |
-
|
192 |
-
|
193 |
-
|
194 |
-
|
195 |
-
|
196 |
-
|
197 |
-
|
198 |
-
|
199 |
-
|
200 |
-
|
201 |
-
|
202 |
-
|
203 |
-
|
204 |
-
|
205 |
-
|
206 |
-
|
207 |
-
|
208 |
-
|
209 |
-
|
210 |
-
|
211 |
-
|
212 |
-
|
213 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# coding=utf-8
|
2 |
+
# Copyright 2024 Microsoft and the HuggingFace Inc. team. All rights reserved.
|
3 |
+
#
|
4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
5 |
+
# you may not use this file except in compliance with the License.
|
6 |
+
# You may obtain a copy of the License at
|
7 |
+
#
|
8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
9 |
+
#
|
10 |
+
# Unless required by applicable law or agreed to in writing, software
|
11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13 |
+
# See the License for the specific language governing permissions and
|
14 |
+
# limitations under the License.
|
15 |
+
|
16 |
+
""" Phi-3 model configuration"""
|
17 |
+
|
18 |
+
|
19 |
+
from transformers.configuration_utils import PretrainedConfig
|
20 |
+
from transformers.utils import logging
|
21 |
+
|
22 |
+
|
23 |
+
logger = logging.get_logger(__name__)
|
24 |
+
|
25 |
+
PHI3_PRETRAINED_CONFIG_ARCHIVE_MAP = {
|
26 |
+
"microsoft/Phi-3-mini-4k-instruct": "https://huggingface.co/microsoft/Phi-3-mini-4k-instruct/resolve/main/config.json",
|
27 |
+
"microsoft/Phi-3-mini-128k-instruct": "https://huggingface.co/microsoft/Phi-3-mini-128k-instruct/resolve/main/config.json",
|
28 |
+
}
|
29 |
+
|
30 |
+
|
31 |
+
class Phi3Config(PretrainedConfig):
|
32 |
+
r"""
|
33 |
+
This is the configuration class to store the configuration of a [`Phi3Model`]. It is used to instantiate a Phi-3
|
34 |
+
model according to the specified arguments, defining the model architecture. Instantiating a configuration with the
|
35 |
+
defaults will yield a similar configuration to that of the
|
36 |
+
[microsoft/Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct).
|
37 |
+
|
38 |
+
Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
|
39 |
+
documentation from [`PretrainedConfig`] for more information.
|
40 |
+
|
41 |
+
Args:
|
42 |
+
vocab_size (`int`, *optional*, defaults to 32064):
|
43 |
+
Vocabulary size of the Phi-3 model. Defines the number of different tokens that can be represented by the
|
44 |
+
`inputs_ids` passed when calling [`Phi3Model`].
|
45 |
+
hidden_size (`int`, *optional*, defaults to 3072):
|
46 |
+
Dimension of the hidden representations.
|
47 |
+
intermediate_size (`int`, *optional*, defaults to 8192):
|
48 |
+
Dimension of the MLP representations.
|
49 |
+
num_hidden_layers (`int`, *optional*, defaults to 32):
|
50 |
+
Number of hidden layers in the Transformer decoder.
|
51 |
+
num_attention_heads (`int`, *optional*, defaults to 32):
|
52 |
+
Number of attention heads for each attention layer in the Transformer decoder.
|
53 |
+
num_key_value_heads (`int`, *optional*):
|
54 |
+
This is the number of key_value heads that should be used to implement Grouped Query Attention. If
|
55 |
+
`num_key_value_heads=num_attention_heads`, the model will use Multi Head Attention (MHA), if
|
56 |
+
`num_key_value_heads=1 the model will use Multi Query Attention (MQA) otherwise GQA is used. When
|
57 |
+
converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed
|
58 |
+
by meanpooling all the original heads within that group. For more details checkout [this
|
59 |
+
paper](https://arxiv.org/pdf/2305.13245.pdf). If it is not specified, will default to
|
60 |
+
`num_attention_heads`.
|
61 |
+
resid_pdrop (`float`, *optional*, defaults to 0.0):
|
62 |
+
Dropout probability for mlp outputs.
|
63 |
+
embd_pdrop (`int`, *optional*, defaults to 0.0):
|
64 |
+
The dropout ratio for the embeddings.
|
65 |
+
attention_dropout (`float`, *optional*, defaults to 0.0):
|
66 |
+
The dropout ratio after computing the attention scores.
|
67 |
+
hidden_act (`str` or `function`, *optional*, defaults to `"silu"`):
|
68 |
+
The non-linear activation function (function or string) in the decoder.
|
69 |
+
max_position_embeddings (`int`, *optional*, defaults to 4096):
|
70 |
+
The maximum sequence length that this model might ever be used with.
|
71 |
+
original_max_position_embeddings (`int`, *optional*, defaults to 4096):
|
72 |
+
The maximum sequence length that this model was trained with. This is used to determine the size of the
|
73 |
+
original RoPE embeddings when using long scaling.
|
74 |
+
initializer_range (`float`, *optional*, defaults to 0.02):
|
75 |
+
The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
|
76 |
+
rms_norm_eps (`float`, *optional*, defaults to 1e-05):
|
77 |
+
The epsilon value used for the RMSNorm.
|
78 |
+
use_cache (`bool`, *optional*, defaults to `True`):
|
79 |
+
Whether or not the model should return the last key/values attentions (not used by all models). Only
|
80 |
+
relevant if `config.is_decoder=True`. Whether to tie weight embeddings or not.
|
81 |
+
tie_word_embeddings (`bool`, *optional*, defaults to `False`):
|
82 |
+
Whether to tie weight embeddings
|
83 |
+
rope_theta (`float`, *optional*, defaults to 10000.0):
|
84 |
+
The base period of the RoPE embeddings.
|
85 |
+
rope_scaling (`dict`, *optional*):
|
86 |
+
The scaling strategy for the RoPE embeddings. If `None`, no scaling is applied. If a dictionary, it must
|
87 |
+
contain the following keys: `type`, `short_factor` and `long_factor`. The `type` must be `longrope` and
|
88 |
+
the `short_factor` and `long_factor` must be lists of numbers with the same length as the hidden size
|
89 |
+
divided by the number of attention heads divided by 2.
|
90 |
+
bos_token_id (`int`, *optional*, defaults to 1):
|
91 |
+
The id of the "beginning-of-sequence" token.
|
92 |
+
eos_token_id (`int`, *optional*, defaults to 32000):
|
93 |
+
The id of the "end-of-sequence" token.
|
94 |
+
pad_token_id (`int`, *optional*, defaults to 32000):
|
95 |
+
The id of the padding token.
|
96 |
+
sliding_window (`int`, *optional*):
|
97 |
+
Sliding window attention window size. If `None`, no sliding window is applied.
|
98 |
+
|
99 |
+
Example:
|
100 |
+
|
101 |
+
```python
|
102 |
+
>>> from transformers import Phi3Model, Phi3Config
|
103 |
+
|
104 |
+
>>> # Initializing a Phi-3 style configuration
|
105 |
+
>>> configuration = Phi3Config.from_pretrained("microsoft/Phi-3-mini-4k-instruct")
|
106 |
+
|
107 |
+
>>> # Initializing a model from the configuration
|
108 |
+
>>> model = Phi3Model(configuration)
|
109 |
+
|
110 |
+
>>> # Accessing the model configuration
|
111 |
+
>>> configuration = model.config
|
112 |
+
```"""
|
113 |
+
|
114 |
+
model_type = "phi3"
|
115 |
+
keys_to_ignore_at_inference = ["past_key_values"]
|
116 |
+
|
117 |
+
def __init__(
|
118 |
+
self,
|
119 |
+
vocab_size=32064,
|
120 |
+
hidden_size=3072,
|
121 |
+
intermediate_size=8192,
|
122 |
+
num_hidden_layers=32,
|
123 |
+
num_attention_heads=32,
|
124 |
+
num_key_value_heads=None,
|
125 |
+
resid_pdrop=0.0,
|
126 |
+
embd_pdrop=0.0,
|
127 |
+
attention_dropout=0.0,
|
128 |
+
hidden_act="silu",
|
129 |
+
max_position_embeddings=4096,
|
130 |
+
original_max_position_embeddings=4096,
|
131 |
+
initializer_range=0.02,
|
132 |
+
rms_norm_eps=1e-5,
|
133 |
+
use_cache=True,
|
134 |
+
tie_word_embeddings=False,
|
135 |
+
rope_theta=10000.0,
|
136 |
+
rope_scaling=None,
|
137 |
+
bos_token_id=1,
|
138 |
+
eos_token_id=32000,
|
139 |
+
pad_token_id=32000,
|
140 |
+
sliding_window=None,
|
141 |
+
**kwargs,
|
142 |
+
):
|
143 |
+
self.vocab_size = vocab_size
|
144 |
+
self.hidden_size = hidden_size
|
145 |
+
self.intermediate_size = intermediate_size
|
146 |
+
self.num_hidden_layers = num_hidden_layers
|
147 |
+
self.num_attention_heads = num_attention_heads
|
148 |
+
|
149 |
+
if num_key_value_heads is None:
|
150 |
+
num_key_value_heads = num_attention_heads
|
151 |
+
|
152 |
+
self.num_key_value_heads = num_key_value_heads
|
153 |
+
self.resid_pdrop = resid_pdrop
|
154 |
+
self.embd_pdrop = embd_pdrop
|
155 |
+
self.attention_dropout = attention_dropout
|
156 |
+
self.hidden_act = hidden_act
|
157 |
+
self.max_position_embeddings = max_position_embeddings
|
158 |
+
self.original_max_position_embeddings = original_max_position_embeddings
|
159 |
+
self.initializer_range = initializer_range
|
160 |
+
self.rms_norm_eps = rms_norm_eps
|
161 |
+
self.use_cache = use_cache
|
162 |
+
self.rope_theta = rope_theta
|
163 |
+
self.rope_scaling = rope_scaling
|
164 |
+
self._rope_scaling_adjustment()
|
165 |
+
self._rope_scaling_validation()
|
166 |
+
self.sliding_window = sliding_window
|
167 |
+
|
168 |
+
super().__init__(
|
169 |
+
bos_token_id=bos_token_id,
|
170 |
+
eos_token_id=eos_token_id,
|
171 |
+
pad_token_id=pad_token_id,
|
172 |
+
tie_word_embeddings=tie_word_embeddings,
|
173 |
+
**kwargs,
|
174 |
+
)
|
175 |
+
|
176 |
+
def _rope_scaling_adjustment(self):
|
177 |
+
"""
|
178 |
+
Adjust the `type` of the `rope_scaling` configuration for backward compatibility.
|
179 |
+
"""
|
180 |
+
if self.rope_scaling is None:
|
181 |
+
return
|
182 |
+
|
183 |
+
rope_scaling_type = self.rope_scaling.get("type", None)
|
184 |
+
|
185 |
+
# For backward compatibility if previous version used "su" or "yarn"
|
186 |
+
if rope_scaling_type is not None and rope_scaling_type in ["su", "yarn"]:
|
187 |
+
self.rope_scaling["type"] = "longrope"
|
188 |
+
|
189 |
+
def _rope_scaling_validation(self):
|
190 |
+
"""
|
191 |
+
Validate the `rope_scaling` configuration.
|
192 |
+
"""
|
193 |
+
if self.rope_scaling is None:
|
194 |
+
return
|
195 |
+
|
196 |
+
if not isinstance(self.rope_scaling, dict) or len(self.rope_scaling) != 3:
|
197 |
+
raise ValueError(
|
198 |
+
"`rope_scaling` must be a dictionary with three fields, `type`, `short_factor` and `long_factor`, "
|
199 |
+
f"got {self.rope_scaling}"
|
200 |
+
)
|
201 |
+
rope_scaling_type = self.rope_scaling.get("type", None)
|
202 |
+
rope_scaling_short_factor = self.rope_scaling.get("short_factor", None)
|
203 |
+
rope_scaling_long_factor = self.rope_scaling.get("long_factor", None)
|
204 |
+
if rope_scaling_type is None or rope_scaling_type not in ["longrope"]:
|
205 |
+
raise ValueError(f"`rope_scaling`'s type field must be one of ['longrope'], got {rope_scaling_type}")
|
206 |
+
if not (
|
207 |
+
isinstance(rope_scaling_short_factor, list)
|
208 |
+
and all(isinstance(x, (int, float)) for x in rope_scaling_short_factor)
|
209 |
+
):
|
210 |
+
raise ValueError(
|
211 |
+
f"`rope_scaling`'s short_factor field must be a list of numbers, got {rope_scaling_short_factor}"
|
212 |
+
)
|
213 |
+
if not len(rope_scaling_short_factor) == self.hidden_size // self.num_attention_heads // 2:
|
214 |
+
raise ValueError(
|
215 |
+
f"`rope_scaling`'s short_factor field must have length {self.hidden_size // self.num_attention_heads // 2}, got {len(rope_scaling_short_factor)}"
|
216 |
+
)
|
217 |
+
if not (
|
218 |
+
isinstance(rope_scaling_long_factor, list)
|
219 |
+
and all(isinstance(x, (int, float)) for x in rope_scaling_long_factor)
|
220 |
+
):
|
221 |
+
raise ValueError(
|
222 |
+
f"`rope_scaling`'s long_factor field must be a list of numbers, got {rope_scaling_long_factor}"
|
223 |
+
)
|
224 |
+
if not len(rope_scaling_long_factor) == self.hidden_size // self.num_attention_heads // 2:
|
225 |
+
raise ValueError(
|
226 |
+
f"`rope_scaling`'s long_factor field must have length {self.hidden_size // self.num_attention_heads // 2}, got {len(rope_scaling_long_factor)}"
|
227 |
+
)
|
model-00001-of-00002.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 5356281360
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:116ff9bda1d393660238f360909bf52758633a0096f26310ef0f522cc41d4c68
|
3 |
size 5356281360
|
model-00002-of-00002.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 2285900500
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b18d25709caa69a4028471c23f922cbce133096419a345371ecfae4289d6eb8f
|
3 |
size 2285900500
|
modeling_phi3.py
CHANGED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer.json
CHANGED
@@ -150,12 +150,6 @@
|
|
150 |
"post_processor": {
|
151 |
"type": "TemplateProcessing",
|
152 |
"single": [
|
153 |
-
{
|
154 |
-
"SpecialToken": {
|
155 |
-
"id": "<s>",
|
156 |
-
"type_id": 0
|
157 |
-
}
|
158 |
-
},
|
159 |
{
|
160 |
"Sequence": {
|
161 |
"id": "A",
|
@@ -164,24 +158,12 @@
|
|
164 |
}
|
165 |
],
|
166 |
"pair": [
|
167 |
-
{
|
168 |
-
"SpecialToken": {
|
169 |
-
"id": "<s>",
|
170 |
-
"type_id": 0
|
171 |
-
}
|
172 |
-
},
|
173 |
{
|
174 |
"Sequence": {
|
175 |
"id": "A",
|
176 |
"type_id": 0
|
177 |
}
|
178 |
},
|
179 |
-
{
|
180 |
-
"SpecialToken": {
|
181 |
-
"id": "<s>",
|
182 |
-
"type_id": 1
|
183 |
-
}
|
184 |
-
},
|
185 |
{
|
186 |
"Sequence": {
|
187 |
"id": "B",
|
@@ -189,17 +171,7 @@
|
|
189 |
}
|
190 |
}
|
191 |
],
|
192 |
-
"special_tokens": {
|
193 |
-
"<s>": {
|
194 |
-
"id": "<s>",
|
195 |
-
"ids": [
|
196 |
-
1
|
197 |
-
],
|
198 |
-
"tokens": [
|
199 |
-
"<s>"
|
200 |
-
]
|
201 |
-
}
|
202 |
-
}
|
203 |
},
|
204 |
"decoder": {
|
205 |
"type": "Sequence",
|
|
|
150 |
"post_processor": {
|
151 |
"type": "TemplateProcessing",
|
152 |
"single": [
|
|
|
|
|
|
|
|
|
|
|
|
|
153 |
{
|
154 |
"Sequence": {
|
155 |
"id": "A",
|
|
|
158 |
}
|
159 |
],
|
160 |
"pair": [
|
|
|
|
|
|
|
|
|
|
|
|
|
161 |
{
|
162 |
"Sequence": {
|
163 |
"id": "A",
|
164 |
"type_id": 0
|
165 |
}
|
166 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
167 |
{
|
168 |
"Sequence": {
|
169 |
"id": "B",
|
|
|
171 |
}
|
172 |
}
|
173 |
],
|
174 |
+
"special_tokens": {}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
175 |
},
|
176 |
"decoder": {
|
177 |
"type": "Sequence",
|
tokenizer_config.json
CHANGED
@@ -1,6 +1,7 @@
|
|
1 |
{
|
2 |
-
"add_bos_token":
|
3 |
"add_eos_token": false,
|
|
|
4 |
"added_tokens_decoder": {
|
5 |
"0": {
|
6 |
"content": "<unk>",
|
@@ -116,9 +117,10 @@
|
|
116 |
}
|
117 |
},
|
118 |
"bos_token": "<s>",
|
119 |
-
"chat_template": "{
|
120 |
"clean_up_tokenization_spaces": false,
|
121 |
"eos_token": "<|endoftext|>",
|
|
|
122 |
"model_max_length": 131072,
|
123 |
"pad_token": "<|endoftext|>",
|
124 |
"padding_side": "left",
|
|
|
1 |
{
|
2 |
+
"add_bos_token": false,
|
3 |
"add_eos_token": false,
|
4 |
+
"add_prefix_space": null,
|
5 |
"added_tokens_decoder": {
|
6 |
"0": {
|
7 |
"content": "<unk>",
|
|
|
117 |
}
|
118 |
},
|
119 |
"bos_token": "<s>",
|
120 |
+
"chat_template": "{% for message in messages %}{% if message['role'] == 'system' %}{{'<|system|>\n' + message['content'] + '<|end|>\n'}}{% elif message['role'] == 'user' %}{{'<|user|>\n' + message['content'] + '<|end|>\n'}}{% elif message['role'] == 'assistant' %}{{'<|assistant|>\n' + message['content'] + '<|end|>\n'}}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '<|assistant|>\n' }}{% else %}{{ eos_token }}{% endif %}",
|
121 |
"clean_up_tokenization_spaces": false,
|
122 |
"eos_token": "<|endoftext|>",
|
123 |
+
"legacy": false,
|
124 |
"model_max_length": 131072,
|
125 |
"pad_token": "<|endoftext|>",
|
126 |
"padding_side": "left",
|