Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly
Sign up
Appearance settings

Commit 1a1f0a5

Browse files
Fix #35, add preset models
1 parent c0b356f commit 1a1f0a5

File tree

1 file changed

+265
-0
lines changed
  • src/main/kotlin/com/cjcrafter/openai

1 file changed

+265
-0
lines changed
Lines changed: 265 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,265 @@
1+
package com.cjcrafter.openai
2+
3+
/**
4+
* Holds all the available models for the OpenAI API. Most users are probably
5+
* interested in [Models.Chat].
6+
*
7+
* Note that this list is manually updated, and may fall out of date. For the
8+
* most updated information, check the [OpenAI documentation](https://platform.openai.com/docs/models).
9+
* If you notice that something is out of date, please [open an issue](https://github.com/CJCrafter/ChatGPT-Java-API/issues).
10+
*
11+
* When OpenAI marks a model as _'Legacy'_, the corresponding field in Models
12+
* will be marked as [Deprecated]. Once it is reported that a model throws an
13+
* error due to deprecation, the deprecation level will be set to [DeprecationLevel.ERROR].
14+
*/
15+
object Models {
16+
17+
/**
18+
* Holds all available Chat models. Chat models are used to generate text
19+
* in a conversational manner. Chat models work using conversational memory.
20+
*
21+
* Note that GPT-4 endpoints are only available to [paying customers](https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4).
22+
*
23+
* @see OpenAI.createChatCompletion
24+
* @see OpenAI.streamChatCompletion
25+
* @see OpenAI.assistants
26+
*/
27+
object Chat {
28+
29+
/**
30+
* `gpt-4` Turbo. Has a context window of 128,000 tokens with training
31+
* data up to April 2023. This model has improved instruction following,
32+
* JSON mode, reproducible output, parallel function calling, and more.
33+
* Returns a maximum of 4,096 output tokens.
34+
*/
35+
const val GPT_4_1106_PREVIEW = "gpt-4-1106-preview"
36+
37+
/**
38+
* `gpt-4` Turbo with vision. Has a context window of 128,000 tokens with
39+
* training data up to April 2023. Has the same capabilities as
40+
* [GPT_4_1106_PREVIEW], but can also understand images.
41+
*/
42+
const val GPT_4_VISION_PREVIEW = "gpt-4-vision-preview"
43+
44+
/**
45+
* Points to the currently supported version of `gpt-4`.
46+
*
47+
* See [continuous model upgrades](https://platform.openai.com/docs/models/continuous-model-upgrades)
48+
*/
49+
const val GPT_4 = "gpt-4"
50+
51+
/**
52+
* Points to the currently supported version of `gpt-4` with a 32k context window.
53+
*
54+
* See [continuous model upgrades](https://platform.openai.com/docs/models/continuous-model-upgrades)
55+
*/
56+
const val GPT_4_32k = "gpt-4-32k"
57+
58+
/**
59+
* Snapshot of `gpt-4` from June 13th, 2023 with improved function calling
60+
* support. Has a context window of 8,192 tokens with training data up
61+
* to September 2021.
62+
*/
63+
const val GPT_4_0613 = "gpt-4-0613"
64+
65+
/**
66+
* Snapshot of `gpt-4-32k` from June 13th, 2023 with improved function
67+
* calling support. Has a context window of 32,768 tokens with training
68+
* data up to September 2021.
69+
*/
70+
const val GPT_4_32k_0613 = "gpt-4-32k-0613"
71+
72+
/**
73+
* Snapshot of `gpt-4` from March 14th 2023 with function calling support.
74+
* This model version will be deprecated on June 13th, 2024. Has a
75+
* context window of 8,192 tokens with training data up to September
76+
* 2021.
77+
*/
78+
@Deprecated(
79+
message = "This model will be removed on June 13th, 2024",
80+
replaceWith = ReplaceWith("GPT_4_0613"),
81+
level = DeprecationLevel.WARNING,
82+
)
83+
const val GPT_4_0314 = "gpt-4-0314"
84+
85+
/**
86+
* Snapshot of `gpt-4` from March 14th 2023 with function calling support.
87+
* This model version will be deprecated on June 13th, 2024. Has a
88+
* context window of 32,768 tokens with training data up to September
89+
* 2021.
90+
*/
91+
@Deprecated(
92+
message = "This model will be removed on June 13th 2024",
93+
replaceWith = ReplaceWith("GPT_4_32k_0613"),
94+
level = DeprecationLevel.WARNING,
95+
)
96+
const val GPT_4_32k_0314 = "gpt-4-32k-0314"
97+
98+
/////////////////////////////////////////////////////
99+
// GPT 3.5 //
100+
/////////////////////////////////////////////////////
101+
102+
/**
103+
* Has a context window of 16,385 tokens with training data up to
104+
* September 2021. This model has improved instruction following, JSON
105+
* mode, reproducible outputs, parallel function calling, and more.
106+
* Returns a maximum of 4,096 output tokens.
107+
*/
108+
const val GPT_3_5_TURBO_1106 = "gpt_3.5-turbo-1106"
109+
110+
/**
111+
* Points to the currently supported version of gpt-3.5-turbo.
112+
*
113+
* See [continuous model upgrades](https://platform.openai.com/docs/models/continuous-model-upgrades)
114+
*/
115+
const val GPT_3_5_TURBO = "gpt-3.5-turbo"
116+
117+
/**
118+
* Points to the currently supported version of gpt-3.5-turbo.
119+
*
120+
* See [continuous model upgrades](https://platform.openai.com/docs/models/continuous-model-upgrades)
121+
*/
122+
const val GPT_3_5_TURBO_16k = "gpt-3.5-turbo-16k"
123+
124+
/**
125+
* Snapshot of `gpt-3.5-turbo` from June 13th, 2023. This model version
126+
* will be deprecated on June 13th, 2024. Has a context window of 4,096
127+
* tokens with training data up to September 2021.
128+
*/
129+
@Deprecated(
130+
message = "This model will be removed on June 13th 2024",
131+
replaceWith = ReplaceWith("GPT_3_5_TURBO_1106"),
132+
level = DeprecationLevel.WARNING,
133+
)
134+
const val GPT_3_5_TURBO_0613 = "gpt-3.5-turbo-0613"
135+
136+
/**
137+
* Snapshot of `gpt-3.5-turbo-16k` from June 13th, 2023. This model
138+
* version will be deprecated on June 13th, 2024. Has a context window
139+
* of 16,385 tokens with training data up to September 2021.
140+
*/
141+
@Deprecated(
142+
message = "This model will be removed on June 13th 2024",
143+
replaceWith = ReplaceWith("GPT_3_5_TURBO_1106"),
144+
level = DeprecationLevel.WARNING,
145+
)
146+
const val GPT_3_5_TURBO_16k_0613 = "gpt-3.5-turbo-16k-0613"
147+
148+
/**
149+
* Snapshot of `gpt-3.5-turbo` from March 1st, 2023. This model version
150+
* will be deprecated on June 13th, 2024. Has a context window of 4,096
151+
* tokens with training data up to September 2021.
152+
*/
153+
@Deprecated(
154+
message = "This model will be removed on June 13th 2024",
155+
replaceWith = ReplaceWith("GPT_3_5_TURBO_1106"),
156+
level = DeprecationLevel.WARNING,
157+
)
158+
const val GPT_3_5_TURBO_0301 = "gpt-3.5-turbo-0301"
159+
}
160+
161+
/**
162+
* Holds all available completion models.
163+
*
164+
* @see OpenAI.createCompletion
165+
* @see OpenAI.streamCompletion
166+
*/
167+
object Completion {
168+
169+
/**
170+
* Similar to `text-davinci-003` but compatible with the legacy
171+
* completions endpoint.
172+
*/
173+
const val GPT_3_5_TURBO_INSTRUCT = "gpt-3.5-turbo-instruct"
174+
175+
/**
176+
* Can do language tasks with better quality and consistency than the
177+
* curie, babbage, or ada models. Will be deprecated on January 4th
178+
* 2024. Has a context window of 4,096 tokens and training data up to
179+
* June 2021.
180+
*/
181+
@Deprecated(
182+
message = "This model will be removed on January 4th 2024",
183+
replaceWith = ReplaceWith("GPT_3_5_TURBO_INSTRUCT"),
184+
level = DeprecationLevel.WARNING,
185+
)
186+
const val TEXT_DAVINCI_003 = "text-davinci-003"
187+
188+
/**
189+
* Similar capabilities to `text-davinci-003` but trained with
190+
* supervised fine-tuning instead of reinforcement learning. Will be
191+
* deprecated on January 4th 2024. Has a context window of 4,096 tokens
192+
* and training data up to June 2021.
193+
*/
194+
@Deprecated(
195+
message = "This model will be removed on January 4th 2024",
196+
replaceWith = ReplaceWith("GPT_3_5_TURBO_INSTRUCT"),
197+
level = DeprecationLevel.WARNING,
198+
)
199+
const val TEXT_DAVINCI_002 = "text-davinci-002"
200+
201+
/**
202+
* Optimized for code-completion tasks. Will be deprecated on Jan 4th
203+
* 2024. Has a context window of 8,001 tokens and training data up to
204+
* June 2021.
205+
*/
206+
@Deprecated(
207+
message = "This model will be removed on January 4th 2024",
208+
replaceWith = ReplaceWith("GPT_3_5_TURBO_INSTRUCT"),
209+
level = DeprecationLevel.WARNING,
210+
)
211+
const val CODE_DAVINCI_002 = "code-davinci-002"
212+
213+
/////////////////////////////////////////////////////
214+
// GPT 3.0 //
215+
// Technically, the models above are part of 3.5 //
216+
/////////////////////////////////////////////////////
217+
218+
/**
219+
* Vary capable, faster and lower cost than `davinci`. Has a context
220+
* window of 2,049 tokens and training data up to October 2019.
221+
*/
222+
const val TEXT_CURIE_001 = "text-curie-001"
223+
224+
/**
225+
* Capable of straightforward tasks, very fast, and lower cost. Has a
226+
* context window of 2,049 tokens and training data up to October 2019.
227+
*/
228+
const val TEXT_BABBAGE_001 = "text-babbage-001"
229+
230+
/**
231+
* Capable of very simple tasks, usually the fastest model in the `gpt-3`
232+
* series, and lowest cost. Has a context window of 2,049 tokens and
233+
* training data up to October 2019.
234+
*/
235+
const val TEXT_ADA_001 = "text-ada-001"
236+
237+
/**
238+
* Most capable `gpt-3` model, can do any task the other models can do,
239+
* often with higher quality. Can be fine-tuned. Has a context window of
240+
* 2,049 tokens and training data up to October 2019.
241+
*/
242+
const val DAVINCI = "davinci"
243+
244+
/**
245+
* Very capable, but faster and lower cost than `davinci`. Can be
246+
* fine-tuned. Has a context window of 2,049 tokens and training
247+
* data up to October 2019.
248+
*/
249+
const val CURIE = "curie"
250+
251+
/**
252+
* Capable of straightforward tasks, very fast, and lower cost. Can be
253+
* fine-tuned. Has a context window of 2,049 tokens and training data
254+
* up to October 2019.
255+
*/
256+
const val BABBAGE = "babbage"
257+
258+
/**
259+
* Capable of very simple tasks, usually the fasted model in the `gpt-3`
260+
* series, and lowest cost. Can be fine-tuned. Has a context window of
261+
* 2,049 tokens and training data up to October 2019.
262+
*/
263+
const val ADA = "ada"
264+
}
265+
}

0 commit comments

Comments
(0)

AltStyle によって変換されたページ (->オリジナル) /