lm-deluge 0.0.34__py3-none-any.whl → 0.0.36__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Potentially problematic release.
This version of lm-deluge might be problematic. Click here for more details.
- lm_deluge/api_requests/anthropic.py +1 -1
- lm_deluge/api_requests/gemini.py +4 -2
- lm_deluge/api_requests/openai.py +17 -4
- lm_deluge/api_requests/response.py +4 -3
- lm_deluge/cli.py +300 -0
- lm_deluge/client.py +22 -73
- lm_deluge/models/__init__.py +144 -0
- lm_deluge/models/anthropic.py +124 -0
- lm_deluge/models/bedrock.py +99 -0
- lm_deluge/models/cerebras.py +57 -0
- lm_deluge/models/cohere.py +98 -0
- lm_deluge/models/deepseek.py +27 -0
- lm_deluge/models/fireworks.py +16 -0
- lm_deluge/models/google.py +153 -0
- lm_deluge/models/grok.py +38 -0
- lm_deluge/models/groq.py +74 -0
- lm_deluge/models/meta.py +65 -0
- lm_deluge/models/mistral.py +110 -0
- lm_deluge/models/openai.py +318 -0
- lm_deluge/models/openrouter.py +1 -0
- lm_deluge/models/together.py +112 -0
- lm_deluge/prompt.py +2 -2
- lm_deluge/util/harmony.py +47 -0
- {lm_deluge-0.0.34.dist-info → lm_deluge-0.0.36.dist-info}/METADATA +1 -1
- {lm_deluge-0.0.34.dist-info → lm_deluge-0.0.36.dist-info}/RECORD +28 -12
- lm_deluge/models.py +0 -1305
- {lm_deluge-0.0.34.dist-info → lm_deluge-0.0.36.dist-info}/WHEEL +0 -0
- {lm_deluge-0.0.34.dist-info → lm_deluge-0.0.36.dist-info}/licenses/LICENSE +0 -0
- {lm_deluge-0.0.34.dist-info → lm_deluge-0.0.36.dist-info}/top_level.txt +0 -0
|
@@ -0,0 +1,318 @@
|
|
|
1
|
+
OPENAI_MODELS = {
|
|
2
|
+
# ███████ █████████ █████
|
|
3
|
+
# ███░░░░░███ ███░░░░░███ ░░███
|
|
4
|
+
# ███ ░░███ ████████ ██████ ████████ ░███ ░███ ░███
|
|
5
|
+
# ░███ ░███░░███░░███ ███░░███░░███░░███ ░███████████ ░███
|
|
6
|
+
# ░███ ░███ ░███ ░███░███████ ░███ ░███ ░███░░░░░███ ░███
|
|
7
|
+
# ░░███ ███ ░███ ░███░███░░░ ░███ ░███ ░███ ░███ ░███
|
|
8
|
+
# ░░░███████░ ░███████ ░░██████ ████ █████ █████ █████ █████
|
|
9
|
+
# ░░░░░░░ ░███░░░ ░░░░░░ ░░░░ ░░░░░ ░░░░░ ░░░░░ ░░░░░
|
|
10
|
+
# ░███
|
|
11
|
+
# █████
|
|
12
|
+
# ░░░░░
|
|
13
|
+
"gpt-5": {
|
|
14
|
+
"id": "gpt-5",
|
|
15
|
+
"name": "gpt-5",
|
|
16
|
+
"api_base": "https://api.openai.com/v1",
|
|
17
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
18
|
+
"supports_json": False,
|
|
19
|
+
"supports_logprobs": True,
|
|
20
|
+
"supports_responses": True,
|
|
21
|
+
"api_spec": "openai",
|
|
22
|
+
"input_cost": 1.25,
|
|
23
|
+
"cached_input_cost": 0.125,
|
|
24
|
+
"output_cost": 10.0,
|
|
25
|
+
"reasoning_model": True,
|
|
26
|
+
},
|
|
27
|
+
"gpt-5-chat": {
|
|
28
|
+
"id": "gpt-5-chat",
|
|
29
|
+
"name": "gpt-5-chat-latest",
|
|
30
|
+
"api_base": "https://api.openai.com/v1",
|
|
31
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
32
|
+
"supports_json": False,
|
|
33
|
+
"supports_logprobs": True,
|
|
34
|
+
"supports_responses": True,
|
|
35
|
+
"api_spec": "openai",
|
|
36
|
+
"input_cost": 1.25,
|
|
37
|
+
"cached_input_cost": 0.125,
|
|
38
|
+
"output_cost": 10.0,
|
|
39
|
+
"reasoning_model": False,
|
|
40
|
+
},
|
|
41
|
+
"gpt-5-mini": {
|
|
42
|
+
"id": "gpt-5-mini",
|
|
43
|
+
"name": "gpt-5-mini",
|
|
44
|
+
"api_base": "https://api.openai.com/v1",
|
|
45
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
46
|
+
"supports_json": False,
|
|
47
|
+
"supports_logprobs": True,
|
|
48
|
+
"supports_responses": True,
|
|
49
|
+
"api_spec": "openai",
|
|
50
|
+
"input_cost": 0.25,
|
|
51
|
+
"cached_input_cost": 0.025,
|
|
52
|
+
"output_cost": 2.0,
|
|
53
|
+
"reasoning_model": True,
|
|
54
|
+
},
|
|
55
|
+
"gpt-5-nano": {
|
|
56
|
+
"id": "gpt-5-nano",
|
|
57
|
+
"name": "gpt-5-nano",
|
|
58
|
+
"api_base": "https://api.openai.com/v1",
|
|
59
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
60
|
+
"supports_json": False,
|
|
61
|
+
"supports_logprobs": True,
|
|
62
|
+
"supports_responses": True,
|
|
63
|
+
"api_spec": "openai",
|
|
64
|
+
"input_cost": 0.05,
|
|
65
|
+
"cached_input_cost": 0.005,
|
|
66
|
+
"output_cost": 0.40,
|
|
67
|
+
"reasoning_model": True,
|
|
68
|
+
},
|
|
69
|
+
"openai-computer-use-preview": {
|
|
70
|
+
"id": "openai-computer-use-preview",
|
|
71
|
+
"name": "computer-use-preview",
|
|
72
|
+
"api_base": "https://api.openai.com/v1",
|
|
73
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
74
|
+
"supports_json": True,
|
|
75
|
+
"supports_logprobs": False,
|
|
76
|
+
"supports_responses": True,
|
|
77
|
+
"api_spec": "openai",
|
|
78
|
+
"input_cost": 2.0,
|
|
79
|
+
"output_cost": 8.0,
|
|
80
|
+
"requests_per_minute": 20,
|
|
81
|
+
"tokens_per_minute": 100_000,
|
|
82
|
+
"reasoning_model": False,
|
|
83
|
+
},
|
|
84
|
+
"o3": {
|
|
85
|
+
"id": "o3",
|
|
86
|
+
"name": "o3-2025-04-16",
|
|
87
|
+
"api_base": "https://api.openai.com/v1",
|
|
88
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
89
|
+
"supports_json": False,
|
|
90
|
+
"supports_logprobs": True,
|
|
91
|
+
"supports_responses": True,
|
|
92
|
+
"api_spec": "openai",
|
|
93
|
+
"input_cost": 10.0,
|
|
94
|
+
"output_cost": 40.0,
|
|
95
|
+
"requests_per_minute": 20,
|
|
96
|
+
"tokens_per_minute": 100_000,
|
|
97
|
+
"reasoning_model": True,
|
|
98
|
+
},
|
|
99
|
+
"o4-mini": {
|
|
100
|
+
"id": "o4-mini",
|
|
101
|
+
"name": "o4-mini-2025-04-16",
|
|
102
|
+
"api_base": "https://api.openai.com/v1",
|
|
103
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
104
|
+
"supports_json": False,
|
|
105
|
+
"supports_logprobs": True,
|
|
106
|
+
"supports_responses": True,
|
|
107
|
+
"api_spec": "openai",
|
|
108
|
+
"input_cost": 1.1,
|
|
109
|
+
"output_cost": 4.4,
|
|
110
|
+
"requests_per_minute": 20,
|
|
111
|
+
"tokens_per_minute": 100_000,
|
|
112
|
+
"reasoning_model": True,
|
|
113
|
+
},
|
|
114
|
+
"gpt-4.1": {
|
|
115
|
+
"id": "gpt-4.1",
|
|
116
|
+
"name": "gpt-4.1-2025-04-14",
|
|
117
|
+
"api_base": "https://api.openai.com/v1",
|
|
118
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
119
|
+
"supports_json": True,
|
|
120
|
+
"supports_logprobs": True,
|
|
121
|
+
"supports_responses": True,
|
|
122
|
+
"api_spec": "openai",
|
|
123
|
+
"input_cost": 2.0,
|
|
124
|
+
"output_cost": 8.0,
|
|
125
|
+
"requests_per_minute": 20,
|
|
126
|
+
"tokens_per_minute": 100_000,
|
|
127
|
+
"reasoning_model": False,
|
|
128
|
+
},
|
|
129
|
+
"gpt-4.1-mini": {
|
|
130
|
+
"id": "gpt-4.1-mini",
|
|
131
|
+
"name": "gpt-4.1-mini-2025-04-14",
|
|
132
|
+
"api_base": "https://api.openai.com/v1",
|
|
133
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
134
|
+
"supports_json": True,
|
|
135
|
+
"supports_logprobs": True,
|
|
136
|
+
"supports_responses": True,
|
|
137
|
+
"api_spec": "openai",
|
|
138
|
+
"input_cost": 0.4,
|
|
139
|
+
"output_cost": 1.6,
|
|
140
|
+
"requests_per_minute": 20,
|
|
141
|
+
"tokens_per_minute": 100_000,
|
|
142
|
+
"reasoning_model": False,
|
|
143
|
+
},
|
|
144
|
+
"gpt-4.1-nano": {
|
|
145
|
+
"id": "gpt-4.1-nano",
|
|
146
|
+
"name": "gpt-4.1-nano-2025-04-14",
|
|
147
|
+
"api_base": "https://api.openai.com/v1",
|
|
148
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
149
|
+
"supports_json": True,
|
|
150
|
+
"supports_logprobs": True,
|
|
151
|
+
"supports_responses": True,
|
|
152
|
+
"api_spec": "openai",
|
|
153
|
+
"input_cost": 0.1,
|
|
154
|
+
"output_cost": 0.4,
|
|
155
|
+
"requests_per_minute": 20,
|
|
156
|
+
"tokens_per_minute": 100_000,
|
|
157
|
+
"reasoning_model": False,
|
|
158
|
+
},
|
|
159
|
+
"gpt-4.5": {
|
|
160
|
+
"id": "gpt-4.5",
|
|
161
|
+
"name": "gpt-4.5-preview-2025-02-27",
|
|
162
|
+
"api_base": "https://api.openai.com/v1",
|
|
163
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
164
|
+
"supports_json": False,
|
|
165
|
+
"supports_logprobs": True,
|
|
166
|
+
"supports_responses": True,
|
|
167
|
+
"api_spec": "openai",
|
|
168
|
+
"input_cost": 75.0,
|
|
169
|
+
"output_cost": 150.0,
|
|
170
|
+
"requests_per_minute": 20,
|
|
171
|
+
"tokens_per_minute": 100_000,
|
|
172
|
+
"reasoning_model": False,
|
|
173
|
+
},
|
|
174
|
+
"o3-mini": {
|
|
175
|
+
"id": "o3-mini",
|
|
176
|
+
"name": "o3-mini-2025-01-31",
|
|
177
|
+
"api_base": "https://api.openai.com/v1",
|
|
178
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
179
|
+
"supports_json": False,
|
|
180
|
+
"supports_logprobs": True,
|
|
181
|
+
"supports_responses": True,
|
|
182
|
+
"api_spec": "openai",
|
|
183
|
+
"input_cost": 1.1,
|
|
184
|
+
"output_cost": 4.4,
|
|
185
|
+
"requests_per_minute": 20,
|
|
186
|
+
"tokens_per_minute": 100_000,
|
|
187
|
+
"reasoning_model": True,
|
|
188
|
+
},
|
|
189
|
+
"o1": {
|
|
190
|
+
"id": "o1",
|
|
191
|
+
"name": "o1-2024-12-17",
|
|
192
|
+
"api_base": "https://api.openai.com/v1",
|
|
193
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
194
|
+
"supports_json": False,
|
|
195
|
+
"supports_logprobs": True,
|
|
196
|
+
"supports_responses": True,
|
|
197
|
+
"api_spec": "openai",
|
|
198
|
+
"input_cost": 15.0,
|
|
199
|
+
"output_cost": 60.0,
|
|
200
|
+
"requests_per_minute": 20,
|
|
201
|
+
"tokens_per_minute": 100_000,
|
|
202
|
+
"reasoning_model": True,
|
|
203
|
+
},
|
|
204
|
+
"o1-preview": {
|
|
205
|
+
"id": "o1-preview",
|
|
206
|
+
"name": "o1-preview-2024-09-12",
|
|
207
|
+
"api_base": "https://api.openai.com/v1",
|
|
208
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
209
|
+
"supports_json": False,
|
|
210
|
+
"supports_logprobs": True,
|
|
211
|
+
"supports_responses": True,
|
|
212
|
+
"api_spec": "openai",
|
|
213
|
+
"input_cost": 15.0,
|
|
214
|
+
"output_cost": 60.0,
|
|
215
|
+
"requests_per_minute": 20,
|
|
216
|
+
"tokens_per_minute": 100_000,
|
|
217
|
+
"reasoning_model": True,
|
|
218
|
+
},
|
|
219
|
+
"o1-mini": {
|
|
220
|
+
"id": "o1-mini",
|
|
221
|
+
"name": "o1-mini-2024-09-12",
|
|
222
|
+
"api_base": "https://api.openai.com/v1",
|
|
223
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
224
|
+
"supports_json": False,
|
|
225
|
+
"supports_logprobs": True,
|
|
226
|
+
"supports_responses": True,
|
|
227
|
+
"api_spec": "openai",
|
|
228
|
+
"input_cost": 3.0,
|
|
229
|
+
"output_cost": 15.0,
|
|
230
|
+
"requests_per_minute": 20,
|
|
231
|
+
"tokens_per_minute": 100_000,
|
|
232
|
+
"reasoning_model": True,
|
|
233
|
+
},
|
|
234
|
+
"gpt-4o": {
|
|
235
|
+
"id": "gpt-4o",
|
|
236
|
+
"name": "gpt-4o-2024-08-06",
|
|
237
|
+
"api_base": "https://api.openai.com/v1",
|
|
238
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
239
|
+
"supports_json": True,
|
|
240
|
+
"supports_logprobs": True,
|
|
241
|
+
"supports_responses": True,
|
|
242
|
+
"api_spec": "openai",
|
|
243
|
+
"input_cost": 5.0,
|
|
244
|
+
"output_cost": 15.0,
|
|
245
|
+
"requests_per_minute": 10_000,
|
|
246
|
+
"tokens_per_minute": 30_000_000,
|
|
247
|
+
},
|
|
248
|
+
"gpt-4o-mini": {
|
|
249
|
+
"id": "gpt-4o-mini",
|
|
250
|
+
"name": "gpt-4o-mini-2024-07-18",
|
|
251
|
+
"api_base": "https://api.openai.com/v1",
|
|
252
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
253
|
+
"supports_json": True,
|
|
254
|
+
"supports_logprobs": True,
|
|
255
|
+
"supports_responses": True,
|
|
256
|
+
"api_spec": "openai",
|
|
257
|
+
"input_cost": 0.15,
|
|
258
|
+
"output_cost": 0.6,
|
|
259
|
+
"requests_per_minute": 60_000,
|
|
260
|
+
"tokens_per_minute": 250_000_000,
|
|
261
|
+
},
|
|
262
|
+
"gpt-3.5-turbo": {
|
|
263
|
+
"id": "gpt-3.5-turbo",
|
|
264
|
+
"name": "gpt-3.5-turbo-0125",
|
|
265
|
+
"api_base": "https://api.openai.com/v1",
|
|
266
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
267
|
+
"supports_json": True,
|
|
268
|
+
"supports_logprobs": True,
|
|
269
|
+
"supports_responses": True,
|
|
270
|
+
"api_spec": "openai",
|
|
271
|
+
"input_cost": 0.5,
|
|
272
|
+
"output_cost": 1.5,
|
|
273
|
+
"requests_per_minute": 40_000,
|
|
274
|
+
"tokens_per_minute": 75_000_000,
|
|
275
|
+
},
|
|
276
|
+
"gpt-4-turbo": {
|
|
277
|
+
"id": "gpt-4-turbo",
|
|
278
|
+
"name": "gpt-4-turbo-2024-04-09",
|
|
279
|
+
"api_base": "https://api.openai.com/v1",
|
|
280
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
281
|
+
"supports_json": True,
|
|
282
|
+
"supports_logprobs": True,
|
|
283
|
+
"supports_responses": True,
|
|
284
|
+
"api_spec": "openai",
|
|
285
|
+
"input_cost": 10.0,
|
|
286
|
+
"output_cost": 30.0,
|
|
287
|
+
"requests_per_minute": 10_000,
|
|
288
|
+
"tokens_per_minute": 1_500_000,
|
|
289
|
+
},
|
|
290
|
+
"gpt-4": {
|
|
291
|
+
"id": "gpt-4",
|
|
292
|
+
"name": "gpt-4-0613",
|
|
293
|
+
"api_base": "https://api.openai.com/v1",
|
|
294
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
295
|
+
"supports_json": False,
|
|
296
|
+
"supports_logprobs": False,
|
|
297
|
+
"supports_responses": True,
|
|
298
|
+
"api_spec": "openai",
|
|
299
|
+
"input_cost": 30.0,
|
|
300
|
+
"output_cost": 60.0,
|
|
301
|
+
"requests_per_minute": 10_000,
|
|
302
|
+
"tokens_per_minute": 300_000,
|
|
303
|
+
},
|
|
304
|
+
"gpt-4-32k": {
|
|
305
|
+
"id": "gpt-4-32k",
|
|
306
|
+
"name": "gpt-4-32k-0613",
|
|
307
|
+
"api_base": "https://api.openai.com/v1",
|
|
308
|
+
"api_key_env_var": "OPENAI_API_KEY",
|
|
309
|
+
"supports_json": False,
|
|
310
|
+
"supports_logprobs": False,
|
|
311
|
+
"supports_responses": True,
|
|
312
|
+
"api_spec": "openai",
|
|
313
|
+
"input_cost": 60.0,
|
|
314
|
+
"output_cost": 120.0,
|
|
315
|
+
"requests_per_minute": 1_000,
|
|
316
|
+
"tokens_per_minute": 150_000,
|
|
317
|
+
},
|
|
318
|
+
}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
OPENROUTER_MODELS = {}
|
|
@@ -0,0 +1,112 @@
|
|
|
1
|
+
# ███████████ █████ █████
|
|
2
|
+
# ░█░░░███░░░█ ░░███ ░░███
|
|
3
|
+
# ░ ░███ ░ ██████ ███████ ██████ ███████ ░███████ ██████ ████████
|
|
4
|
+
# ░███ ███░░███ ███░░███ ███░░███░░░███░ ░███░░███ ███░░███░░███░░███
|
|
5
|
+
# ░███ ░███ ░███░███ ░███░███████ ░███ ░███ ░███ ░███████ ░███ ░░░
|
|
6
|
+
# ░███ ░███ ░███░███ ░███░███░░░ ░███ ███ ░███ ░███ ░███░░░ ░███
|
|
7
|
+
# █████ ░░██████ ░░███████░░██████ ░░█████ ████ █████░░██████ █████
|
|
8
|
+
# ░░░░░ ░░░░░░ ░░░░░███ ░░░░░░ ░░░░░ ░░░░ ░░░░░ ░░░░░░ ░░░░░
|
|
9
|
+
# ███ ░███
|
|
10
|
+
# ░░██████
|
|
11
|
+
# ░░░░░░
|
|
12
|
+
# tbh only reason to use these are that they're cheap, but all worse than haiku
|
|
13
|
+
TOGETHER_MODELS = {
|
|
14
|
+
"deepseek-r1-together": {
|
|
15
|
+
"id": "deepseek-r1-together",
|
|
16
|
+
"name": "deepseek-ai/DeepSeek-R1",
|
|
17
|
+
"api_base": "https://api.together.xyz/v1",
|
|
18
|
+
"api_key_env_var": "TOGETHER_API_KEY",
|
|
19
|
+
"supports_json": False,
|
|
20
|
+
"api_spec": "openai",
|
|
21
|
+
"input_cost": 3.0,
|
|
22
|
+
"output_cost": 7.0,
|
|
23
|
+
"requests_per_minute": None,
|
|
24
|
+
"tokens_per_minute": None,
|
|
25
|
+
},
|
|
26
|
+
"deepseek-v3-together": {
|
|
27
|
+
"id": "deepseek-v3-together",
|
|
28
|
+
"name": "deepseek-ai/DeepSeek-V3",
|
|
29
|
+
"api_base": "https://api.together.xyz/v1",
|
|
30
|
+
"api_key_env_var": "TOGETHER_API_KEY",
|
|
31
|
+
"supports_json": False,
|
|
32
|
+
"api_spec": "openai",
|
|
33
|
+
"input_cost": 1.25,
|
|
34
|
+
"output_cost": 1.25,
|
|
35
|
+
"requests_per_minute": None,
|
|
36
|
+
"tokens_per_minute": None,
|
|
37
|
+
},
|
|
38
|
+
"qwen-3-235b-together": {
|
|
39
|
+
"id": "qwen-3-235b-together",
|
|
40
|
+
"name": "Qwen/Qwen3-235B-A22B-fp8",
|
|
41
|
+
"api_base": "https://api.together.xyz/v1",
|
|
42
|
+
"api_key_env_var": "TOGETHER_API_KEY",
|
|
43
|
+
"supports_json": False,
|
|
44
|
+
"api_spec": "openai",
|
|
45
|
+
"input_cost": 0.2,
|
|
46
|
+
"output_cost": 0.6,
|
|
47
|
+
"requests_per_minute": None,
|
|
48
|
+
"tokens_per_minute": None,
|
|
49
|
+
},
|
|
50
|
+
"qwen-2.5-vl-together": {
|
|
51
|
+
"id": "qwen-2.5-vl-together",
|
|
52
|
+
"name": "Qwen/Qwen2.5-VL-72B-Instruct",
|
|
53
|
+
"api_base": "https://api.together.xyz/v1",
|
|
54
|
+
"api_key_env_var": "TOGETHER_API_KEY",
|
|
55
|
+
"supports_json": False,
|
|
56
|
+
"api_spec": "openai",
|
|
57
|
+
"input_cost": 1.95,
|
|
58
|
+
"output_cost": 8.0,
|
|
59
|
+
"requests_per_minute": None,
|
|
60
|
+
"tokens_per_minute": None,
|
|
61
|
+
},
|
|
62
|
+
"llama-4-maverick-together": {
|
|
63
|
+
"id": "llama-4-maverick-together",
|
|
64
|
+
"name": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8",
|
|
65
|
+
"api_base": "https://api.together.xyz/v1",
|
|
66
|
+
"api_key_env_var": "TOGETHER_API_KEY",
|
|
67
|
+
"supports_json": False,
|
|
68
|
+
"api_spec": "openai",
|
|
69
|
+
"input_cost": 0.27,
|
|
70
|
+
"output_cost": 0.85,
|
|
71
|
+
"requests_per_minute": None,
|
|
72
|
+
"tokens_per_minute": None,
|
|
73
|
+
},
|
|
74
|
+
"llama-4-scout-together": {
|
|
75
|
+
"id": "llama-4-scout-together",
|
|
76
|
+
"name": "meta-llama/Llama-4-Scout-17B-16E-Instruct",
|
|
77
|
+
"api_base": "https://api.together.xyz/v1",
|
|
78
|
+
"api_key_env_var": "TOGETHER_API_KEY",
|
|
79
|
+
"supports_json": False,
|
|
80
|
+
"api_spec": "openai",
|
|
81
|
+
"input_cost": 0.18,
|
|
82
|
+
"output_cost": 0.59,
|
|
83
|
+
"requests_per_minute": None,
|
|
84
|
+
"tokens_per_minute": None,
|
|
85
|
+
},
|
|
86
|
+
"gpt-oss-120b-together": {
|
|
87
|
+
"id": "gpt-oss-120b-together",
|
|
88
|
+
"name": "openai/gpt-oss-120b",
|
|
89
|
+
"api_base": "https://api.together.xyz/v1",
|
|
90
|
+
"api_key_env_var": "TOGETHER_API_KEY",
|
|
91
|
+
"supports_json": False,
|
|
92
|
+
"api_spec": "openai",
|
|
93
|
+
"input_cost": 0.18,
|
|
94
|
+
"output_cost": 0.59,
|
|
95
|
+
"requests_per_minute": None,
|
|
96
|
+
"tokens_per_minute": None,
|
|
97
|
+
"reasoning_model": True,
|
|
98
|
+
},
|
|
99
|
+
"gpt-oss-20b-together": {
|
|
100
|
+
"id": "gpt-oss-20b-together",
|
|
101
|
+
"name": "openai/gpt-oss-20b",
|
|
102
|
+
"api_base": "https://api.together.xyz/v1",
|
|
103
|
+
"api_key_env_var": "TOGETHER_API_KEY",
|
|
104
|
+
"supports_json": False,
|
|
105
|
+
"api_spec": "openai",
|
|
106
|
+
"input_cost": 0.18,
|
|
107
|
+
"output_cost": 0.59,
|
|
108
|
+
"requests_per_minute": None,
|
|
109
|
+
"tokens_per_minute": None,
|
|
110
|
+
"reasoning_model": True,
|
|
111
|
+
},
|
|
112
|
+
}
|
lm_deluge/prompt.py
CHANGED
|
@@ -654,11 +654,11 @@ class Conversation:
|
|
|
654
654
|
pass
|
|
655
655
|
|
|
656
656
|
# fluent additions
|
|
657
|
-
def
|
|
657
|
+
def with_message(self, msg: Message) -> "Conversation":
|
|
658
658
|
self.messages.append(msg)
|
|
659
659
|
return self
|
|
660
660
|
|
|
661
|
-
def
|
|
661
|
+
def with_tool_result(
|
|
662
662
|
self, tool_call_id: str, result: str | list[ToolResultPart]
|
|
663
663
|
) -> "Conversation":
|
|
664
664
|
"""Add a tool result to the conversation.
|
|
@@ -0,0 +1,47 @@
|
|
|
1
|
+
# sample thing we'd want to parse from llama.cpp
|
|
2
|
+
# the goal here is: barebones inference implementation returns
|
|
3
|
+
# raw harmony string; we parse into content blocks
|
|
4
|
+
|
|
5
|
+
# implied: <|start|>assistant
|
|
6
|
+
# <|channel|>analysis<|message|>We need to respond as a helpful assistant. The user says "who are you and what do you want with my family?" This is a normal question. We should answer that we are ChatGPT, an AI language model, and we don't want anything with their family. We reassure them.<|start|>assistant<|channel|>final<|message|>I’m ChatGPT, a large language‑model AI created by OpenAI. I don’t have personal intentions or desires, and I’m not able to interact with anyone outside of this chat. My only goal here is to provide information, answer questions, and help you with whatever you need—nothing more, nothing less. If you have any concerns or need help with something specific, just let me know!
|
|
7
|
+
#
|
|
8
|
+
import copy
|
|
9
|
+
from lm_deluge.api_requests.response import APIResponse
|
|
10
|
+
from lm_deluge.prompt import Text, Thinking
|
|
11
|
+
|
|
12
|
+
SAMPLE_INPUT = """
|
|
13
|
+
<|channel|>analysis<|message|>We need to respond as a helpful assistant. The user says "who are you and what do you want with my family?" This is a normal question. We should answer that we are ChatGPT, an AI language model, and we don't want anything with their family. We reassure them.<|start|>assistant<|channel|>final<|message|>I’m ChatGPT, a large language‑model AI created by OpenAI. I don’t have personal intentions or desires, and I’m not able to interact with anyone outside of this chat. My only goal here is to provide information, answer questions, and help you with whatever you need—nothing more, nothing less. If you have any concerns or need help with something specific, just let me know!
|
|
14
|
+
""".strip()
|
|
15
|
+
|
|
16
|
+
|
|
17
|
+
def _split_messages(response: str):
|
|
18
|
+
raw_messages = response.split("<|start|>")
|
|
19
|
+
messages = []
|
|
20
|
+
for msg in raw_messages:
|
|
21
|
+
channel, content = msg.split("<|message|>")
|
|
22
|
+
channel = channel.split("<|channel|>")[1]
|
|
23
|
+
messages.append((channel, content))
|
|
24
|
+
|
|
25
|
+
return messages
|
|
26
|
+
|
|
27
|
+
|
|
28
|
+
def postprocess_harmony(response: APIResponse) -> APIResponse:
|
|
29
|
+
if not response.content:
|
|
30
|
+
return response
|
|
31
|
+
|
|
32
|
+
parts = response.content.parts
|
|
33
|
+
assert len(parts) == 1, "expected 1 parts to convert harmony"
|
|
34
|
+
text = parts[0].text # type: ignore
|
|
35
|
+
messages = _split_messages(text)
|
|
36
|
+
|
|
37
|
+
new_parts = []
|
|
38
|
+
for channel, content in messages:
|
|
39
|
+
if channel == "analysis":
|
|
40
|
+
new_parts.append(Thinking(content=content))
|
|
41
|
+
elif channel == "final":
|
|
42
|
+
new_parts.append(Text(text=content))
|
|
43
|
+
|
|
44
|
+
new_response = copy.deepcopy(response)
|
|
45
|
+
new_response.content.parts = new_parts # type: ignore
|
|
46
|
+
|
|
47
|
+
return new_response
|
|
@@ -2,29 +2,29 @@ lm_deluge/__init__.py,sha256=mAztMuxINmh7dGbYnT8tsmw1eryQAvd0jpY8yHzd0EE,315
|
|
|
2
2
|
lm_deluge/agent.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
|
3
3
|
lm_deluge/batches.py,sha256=vJXVnuuGkIQnXoDPODPERrvdG9X1Ov1jnXExnPe6ZAc,21772
|
|
4
4
|
lm_deluge/cache.py,sha256=VB1kv8rM2t5XWPR60uhszFcxLDnVKOe1oA5hYjVDjIo,4375
|
|
5
|
-
lm_deluge/
|
|
5
|
+
lm_deluge/cli.py,sha256=Ilww5gOw3J5v0NReq_Ra4hhxU4BCIJBl1oTGxJZKedc,12065
|
|
6
|
+
lm_deluge/client.py,sha256=YKZB8oJx58n8Q5kLV6hT1HeYgxvZGro5RQVH9idqJMU,32576
|
|
6
7
|
lm_deluge/config.py,sha256=H1tQyJDNHGFuwxqQNL5Z-CjWAC0luHSBA3iY_pxmACM,932
|
|
7
8
|
lm_deluge/embed.py,sha256=CO-TOlC5kOTAM8lcnicoG4u4K664vCBwHF1vHa-nAGg,13382
|
|
8
9
|
lm_deluge/errors.py,sha256=oHjt7YnxWbh-eXMScIzov4NvpJMo0-2r5J6Wh5DQ1tk,209
|
|
9
10
|
lm_deluge/file.py,sha256=FGomcG8s2go_55Z2CChflHgmU-UqgFftgFY8c7f_G70,5631
|
|
10
11
|
lm_deluge/gemini_limits.py,sha256=V9mpS9JtXYz7AY6OuKyQp5TuIMRH1BVv9YrSNmGmHNA,1569
|
|
11
12
|
lm_deluge/image.py,sha256=Qpa0k5yXfrpSaHzVUwW_TEn7yEgmwzYGL17Sa7-KhSA,7729
|
|
12
|
-
lm_deluge/
|
|
13
|
-
lm_deluge/prompt.py,sha256=cfwzCAmT-1K0v7SfEMUrxpBkJGgf7IFlWfNLJrCcoBM,37025
|
|
13
|
+
lm_deluge/prompt.py,sha256=gRGu_9wWWMusM7sf-YCdotcZUt1Cj_h_1_6oyS7XTYM,37035
|
|
14
14
|
lm_deluge/request_context.py,sha256=o33LSEwnK6YPhZeulUoSE_VrdKCXiCQa0tjjixK2K6M,2540
|
|
15
15
|
lm_deluge/rerank.py,sha256=-NBAJdHz9OB-SWWJnHzkFmeVO4wR6lFV7Vw-SxG7aVo,11457
|
|
16
16
|
lm_deluge/tool.py,sha256=_coOKB9nPNVZoseMRumRyQ8BMR7_d0IlstzMHNT69JY,15732
|
|
17
17
|
lm_deluge/tracker.py,sha256=rTOjPEwaNczNz9MKDGayPNdmDZOpIWvLll7uz0CloVU,11533
|
|
18
18
|
lm_deluge/usage.py,sha256=VMEKghePFIID5JFBObqYxFpgYxnbYm_dnHy7V1-_T6M,4866
|
|
19
19
|
lm_deluge/api_requests/__init__.py,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
|
|
20
|
-
lm_deluge/api_requests/anthropic.py,sha256=
|
|
20
|
+
lm_deluge/api_requests/anthropic.py,sha256=J5BzYV7aYNoL6FPArB6usyS267z1BguZTRY5JLMd0So,8159
|
|
21
21
|
lm_deluge/api_requests/base.py,sha256=EVHNFtlttKbN7Tt1MnLaO-NjvKHPSV5CqlRv-OnpVAE,5593
|
|
22
22
|
lm_deluge/api_requests/bedrock.py,sha256=FZMhF590JzJtAYDugbDtG93RhPt5efWZ0Wn4V8U8Dgw,11031
|
|
23
23
|
lm_deluge/api_requests/common.py,sha256=BZ3vRO5TB669_UsNKugkkuFSzoLHOYJIKt4nV4sf4vc,422
|
|
24
|
-
lm_deluge/api_requests/gemini.py,sha256=
|
|
24
|
+
lm_deluge/api_requests/gemini.py,sha256=COHqPWmeaq9fpg0YwOZqQTUbijKnXNF4cvMLnW9kLl8,7857
|
|
25
25
|
lm_deluge/api_requests/mistral.py,sha256=S_LpOfCGbCVEROH_od3P-tYeNYTKFMamMTL-c_wFCBI,4597
|
|
26
|
-
lm_deluge/api_requests/openai.py,sha256=
|
|
27
|
-
lm_deluge/api_requests/response.py,sha256=
|
|
26
|
+
lm_deluge/api_requests/openai.py,sha256=FL_UCELdkaf_GZIBPViLdNcUwPMwqvEKj9mMcH72Nmc,22346
|
|
27
|
+
lm_deluge/api_requests/response.py,sha256=Zc9kxBqB4JJIFR6OhXW-BS3ulK5JygE75JNBEpKgn5Q,5989
|
|
28
28
|
lm_deluge/api_requests/deprecated/bedrock.py,sha256=WrcIShCoO8JCUSlFOCHxg6KQCNTZfw3TpYTvSpYk4mA,11320
|
|
29
29
|
lm_deluge/api_requests/deprecated/cohere.py,sha256=KgDScD6_bWhAzOY5BHZQKSA3kurt4KGENqC4wLsGmcU,5142
|
|
30
30
|
lm_deluge/api_requests/deprecated/deepseek.py,sha256=FEApI93VAWDwuaqTooIyKMgONYqRhdUmiAPBRme-IYs,4582
|
|
@@ -43,13 +43,29 @@ lm_deluge/llm_tools/locate.py,sha256=lYNbKTmy9dTvj0lEQkOQ7yrxyqsgYzjD0C_byJKI_4w
|
|
|
43
43
|
lm_deluge/llm_tools/ocr.py,sha256=7fDlvs6uUOvbxMasvGGNJx5Fj6biM6z3lijKZaGN26k,23
|
|
44
44
|
lm_deluge/llm_tools/score.py,sha256=9oGA3-k2U5buHQXkXaEI9M4Wb5yysNhTLsPbGeghAlQ,2580
|
|
45
45
|
lm_deluge/llm_tools/translate.py,sha256=iXyYvQZ8bC44FWhBk4qpdqjKM1WFF7Shq-H2PxhPgg4,1452
|
|
46
|
+
lm_deluge/models/__init__.py,sha256=cwVidzBhMS6B9zOemDf0rAUHo8rHVum8XRE7z5Gi4F4,4278
|
|
47
|
+
lm_deluge/models/anthropic.py,sha256=3pW7fyBY9Xh1m1RtfncU9amWTtKnjGZD0STjpu8iUSQ,5700
|
|
48
|
+
lm_deluge/models/bedrock.py,sha256=jpb_n-Wh3G3VAKZn7U1t5r5IQ2oTDXwrjGIP013l2cI,4534
|
|
49
|
+
lm_deluge/models/cerebras.py,sha256=5Mp1rLWKRLXKpfk9Ef-ydmcp8ffQlNXbp3Zg8sh1pEs,2017
|
|
50
|
+
lm_deluge/models/cohere.py,sha256=M_7cVA9QD4qe1X4sZXCpKEkKrKz2jibaspiTnzsZ1GU,3998
|
|
51
|
+
lm_deluge/models/deepseek.py,sha256=6_jDEprNNYis5I5MDQNloRes9h1P6pMYHXxOd2UZMgg,941
|
|
52
|
+
lm_deluge/models/fireworks.py,sha256=4D0LUgl1QBISGGA9qC62858glju2RRO0uP6X8QRAa4Y,572
|
|
53
|
+
lm_deluge/models/google.py,sha256=PWKVf6HLt9m_HSTvVavZ8BlriQBEtS47ir3jBvetkaQ,5823
|
|
54
|
+
lm_deluge/models/grok.py,sha256=aInkUSclXE47Lm4PKiP3OebAP9V-GOZwK-Eiis4zVow,1199
|
|
55
|
+
lm_deluge/models/groq.py,sha256=BHuBNUpcjsTpwXbnKVfmZf7oef81U48IymR_isMCzvo,2482
|
|
56
|
+
lm_deluge/models/meta.py,sha256=m6HPR82TJONYTTWkQw5EKmITMxoWzrfYOuNgFnGaRX8,2195
|
|
57
|
+
lm_deluge/models/mistral.py,sha256=x67o5gckBGmPcIGdVbS26XZAYFKBYM4tsxEAahGp8bk,4323
|
|
58
|
+
lm_deluge/models/openai.py,sha256=q3IqHldFJjRz-jxT2NoQW9t1_c_BGLd72d1HZlxXiLA,11100
|
|
59
|
+
lm_deluge/models/openrouter.py,sha256=aAgBT5_TZQtUPQyNn-Bob6NGyrlFOclnxIb0F53pgvA,23
|
|
60
|
+
lm_deluge/models/together.py,sha256=RCZoYAb8OVxdH9uwXnv47TDTGzC30P-FZoDbiBE23_g,4957
|
|
61
|
+
lm_deluge/util/harmony.py,sha256=XBfJck6q-5HbOqMhEjdfy1i17i0QtpHG8ruXV4EsHl0,2731
|
|
46
62
|
lm_deluge/util/json.py,sha256=_4Oar2Cmz2L1DK3EtPLPDxD6rsYHxjROmV8ZpmMjQ-4,5822
|
|
47
63
|
lm_deluge/util/logprobs.py,sha256=UkBZakOxWluaLqHrjARu7xnJ0uCHVfLGHJdnYlEcutk,11768
|
|
48
64
|
lm_deluge/util/spatial.py,sha256=BsF_UKhE-x0xBirc-bV1xSKZRTUhsOBdGqsMKme20C8,4099
|
|
49
65
|
lm_deluge/util/validation.py,sha256=hz5dDb3ebvZrZhnaWxOxbNSVMI6nmaOODBkk0htAUhs,1575
|
|
50
66
|
lm_deluge/util/xml.py,sha256=Ft4zajoYBJR3HHCt2oHwGfymGLdvp_gegVmJ-Wqk4Ck,10547
|
|
51
|
-
lm_deluge-0.0.
|
|
52
|
-
lm_deluge-0.0.
|
|
53
|
-
lm_deluge-0.0.
|
|
54
|
-
lm_deluge-0.0.
|
|
55
|
-
lm_deluge-0.0.
|
|
67
|
+
lm_deluge-0.0.36.dist-info/licenses/LICENSE,sha256=uNNXGXPCw2TC7CUs7SEBkA-Mz6QBQFWUUEWDMgEs1dU,1058
|
|
68
|
+
lm_deluge-0.0.36.dist-info/METADATA,sha256=VB8l79DLczVf9_yr9WZRnxdvw5qq55grpBrhBkz-NUs,13295
|
|
69
|
+
lm_deluge-0.0.36.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
|
|
70
|
+
lm_deluge-0.0.36.dist-info/top_level.txt,sha256=hqU-TJX93yBwpgkDtYcXyLr3t7TLSCCZ_reytJjwBaE,10
|
|
71
|
+
lm_deluge-0.0.36.dist-info/RECORD,,
|