Nocare3 commited on
Commit
1e1669d
·
verified ·
1 Parent(s): 1589b50

Delete distiset_configs

Browse files
distiset_configs/README.md DELETED
@@ -1,73 +0,0 @@
1
- ---
2
- size_categories: n<1K
3
- tags:
4
- - synthetic
5
- - distilabel
6
- - rlaif
7
- ---
8
-
9
- <p align="left">
10
- <a href="https://github.com/argilla-io/distilabel">
11
- <img src="https://raw.githubusercontent.com/argilla-io/distilabel/main/docs/assets/distilabel-badge-light.png" alt="Built with Distilabel" width="200" height="32"/>
12
- </a>
13
- </p>
14
-
15
- # Dataset Card for love2dapi_queries
16
-
17
- This dataset has been created with [distilabel](https://distilabel.argilla.io/).
18
-
19
- ## Dataset Summary
20
-
21
- This dataset contains a `pipeline.yaml` which can be used to reproduce the pipeline that generated it in distilabel using the `distilabel` CLI:
22
-
23
- ```console
24
- distilabel pipeline run --config "https://huggingface.co/datasets/love2dapi_queries/raw/main/pipeline.yaml"
25
- ```
26
-
27
- or explore the configuration:
28
-
29
- ```console
30
- distilabel pipeline info --config "https://huggingface.co/datasets/love2dapi_queries/raw/main/pipeline.yaml"
31
- ```
32
-
33
- ## Dataset structure
34
-
35
- The examples have the following structure per configuration:
36
-
37
-
38
- <details><summary> Configuration: default </summary><hr>
39
-
40
- ```json
41
- {
42
- "anchor": "description: Argilla is a collaboration platform for AI engineers and domain experts that require high-quality outputs, full data ownership, and overall efficiency.\nhide: navigation\n\nWelcome to Argilla\n\nArgilla is a collaboration platform for AI engineers and domain experts that require high-quality outputs, full data ownership, and overall efficiency.",
43
- "distilabel_metadata": {
44
- "raw_output_multiply_queries": null
45
- },
46
- "filename": "argilla-python/docs/index.md",
47
- "model_name_query": "meta-llama/Meta-Llama-3-70B-Instruct",
48
- "model_name_query_multiplied": "meta-llama/Meta-Llama-3-70B-Instruct",
49
- "negative": null,
50
- "positive": null,
51
- "queries": null,
52
- "repo_name": "argilla-io/argilla-python"
53
- }
54
- ```
55
-
56
- This subset can be loaded as:
57
-
58
- ```python
59
- from datasets import load_dataset
60
-
61
- ds = load_dataset("love2dapi_queries", "default")
62
- ```
63
-
64
- Or simply as it follows, since there's only one configuration and is named `default`:
65
-
66
- ```python
67
- from datasets import load_dataset
68
-
69
- ds = load_dataset("love2dapi_queries")
70
- ```
71
-
72
-
73
- </details>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
distiset_configs/pipeline.log DELETED
@@ -1,38 +0,0 @@
1
- [2024-07-19 16:51:10] WARNING Since the `base_url=https://api-inference.huggingface.co/models/meta-llama/Meta-Llama-3-70B-Instruct` is available and either one of `model_id` or `endpoint_name` is also provided, the `base_url` will either be ignored or overwritten with the one generated from either of those args, for serverless or dedicated inference endpoints, respectively.
2
- [2024-07-19 16:51:10] WARNING Since the `base_url=https://api-inference.huggingface.co/models/meta-llama/Meta-Llama-3-70B-Instruct` is available and either one of `model_id` or `endpoint_name` is also provided, the `base_url` will either be ignored or overwritten with the one generated from either of those args, for serverless or dedicated inference endpoints, respectively.
3
- [2024-07-19 16:53:15] WARNING Since the `base_url=https://api-inference.huggingface.co/models/meta-llama/Meta-Llama-3-70B-Instruct` is available and either one of `model_id` or `endpoint_name` is also provided, the `base_url` will either be ignored or overwritten with the one generated from either of those args, for serverless or dedicated inference endpoints, respectively.
4
- [2024-07-19 16:53:15] WARNING Since the `base_url=https://api-inference.huggingface.co/models/meta-llama/Meta-Llama-3-70B-Instruct` is available and either one of `model_id` or `endpoint_name` is also provided, the `base_url` will either be ignored or overwritten with the one generated from either of those args, for serverless or dedicated inference endpoints, respectively.
5
- [2024-07-19 17:03:37] WARNING Since the `base_url=https://api-inference.huggingface.co/models/meta-llama/Meta-Llama-3-70B-Instruct` is available and either one of `model_id` or `endpoint_name` is also provided, the `base_url` will either be ignored or overwritten with the one generated from either of those args, for serverless or dedicated inference endpoints, respectively.
6
- [2024-07-19 17:03:37] WARNING Since the `base_url=https://api-inference.huggingface.co/models/meta-llama/Meta-Llama-3-70B-Instruct` is available and either one of `model_id` or `endpoint_name` is also provided, the `base_url` will either be ignored or overwritten with the one generated from either of those args, for serverless or dedicated inference endpoints, respectively.
7
- [2024-07-19 17:03:52] WARNING Task 'multiply_queries' failed to format output: 'NoneType' object has no attribute 'split'. Saving raw response.
8
- [2024-07-19 17:03:52] WARNING Subprocess traceback:
9
-
10
- Traceback (most recent call last):
11
- File "C:\Users\Andi\Python projects\RAGTesting\.venv\Lib\site-packages\distilabel\pipeline\local.py", line 512, in _non_generator_process_loop
12
- result = next(self.step.process_applying_mappings(*batch.data))
13
- ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
14
- File "C:\Users\Andi\Python projects\RAGTesting\.venv\Lib\site-packages\distilabel\steps\base.py", line 512, in process_applying_mappings
15
- for output_rows in generator:
16
- File "C:\Users\Andi\Python projects\RAGTesting\.venv\Lib\site-packages\distilabel\steps\combine.py", line 119, in process
17
- yield combine_dicts(
18
- ^^^^^^^^^^^^^^
19
- File "C:\Users\Andi\Python projects\RAGTesting\.venv\Lib\site-packages\distilabel\pipeline\utils.py", line 39, in combine_dicts
20
- raise ValueError(
21
- ValueError: The length of output_merge_keys must be the same as the length of merge_keys
22
-
23
- [2024-07-19 17:03:52] WARNING Subprocess traceback:
24
-
25
- Traceback (most recent call last):
26
- File "C:\Users\Andi\Python projects\RAGTesting\.venv\Lib\site-packages\distilabel\pipeline\local.py", line 512, in _non_generator_process_loop
27
- result = next(self.step.process_applying_mappings(*batch.data))
28
- ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
29
- File "C:\Users\Andi\Python projects\RAGTesting\.venv\Lib\site-packages\distilabel\steps\base.py", line 512, in process_applying_mappings
30
- for output_rows in generator:
31
- File "C:\Users\Andi\Python projects\RAGTesting\.venv\Lib\site-packages\distilabel\steps\expand.py", line 111, in process
32
- yield [row for input in inputs for row in self._expand_columns(input)]
33
- ^^^^^^^^^^^^^^^^^^^^^^^^^^^
34
- File "C:\Users\Andi\Python projects\RAGTesting\.venv\Lib\site-packages\distilabel\steps\expand.py", line 126, in _expand_columns
35
- for item, expanded in zip_longest(*[data, expanded_rows], fillvalue=input):
36
- ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
37
- TypeError: 'NoneType' object is not iterable
38
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
distiset_configs/pipeline.yaml DELETED
@@ -1,334 +0,0 @@
1
- distilabel:
2
- version: 1.2.2
3
- pipeline:
4
- connections:
5
- - from: load_data
6
- to:
7
- - generate_sentence_pair
8
- - from: generate_sentence_pair
9
- to:
10
- - multiply_queries
11
- - from: multiply_queries
12
- to:
13
- - merge_columns
14
- - from: merge_columns
15
- to:
16
- - expand_columns_0
17
- - from: expand_columns_0
18
- to: []
19
- description: Generate queries to train a sentence embedding model.
20
- name: embedding-queries
21
- routing_batch_functions: []
22
- steps:
23
- - name: load_data
24
- step:
25
- batch_size: 10
26
- config: null
27
- input_mappings: {}
28
- name: load_data
29
- num_examples: null
30
- output_mappings:
31
- chunks: anchor
32
- repo_id: Nocare3/love2dapi_chunks
33
- runtime_parameters_info:
34
- - description: The number of rows that will contain the batches generated by
35
- the step.
36
- name: batch_size
37
- optional: true
38
- - description: The Hugging Face Hub repository ID of the dataset to load.
39
- name: repo_id
40
- optional: false
41
- - description: The split of the dataset to load. Defaults to 'train'.
42
- name: split
43
- optional: true
44
- - description: The configuration of the dataset to load. This is optional and
45
- only needed if the dataset has multiple configurations.
46
- name: config
47
- optional: true
48
- - description: Whether to load the dataset in streaming mode or not. Defaults
49
- to False.
50
- name: streaming
51
- optional: true
52
- - description: The number of examples to load from the dataset. By default will
53
- load all examples.
54
- name: num_examples
55
- optional: true
56
- split: train
57
- storage_options: null
58
- streaming: false
59
- type_info:
60
- module: distilabel.steps.generators.huggingface
61
- name: LoadDataFromHub
62
- - name: generate_sentence_pair
63
- step:
64
- action: query
65
- add_raw_output: true
66
- context: The generated sentence has to be related with Love2d, a lua-code game
67
- engine used mostly by indie developers.
68
- group_generations: false
69
- input_batch_size: 10
70
- input_mappings: {}
71
- llm:
72
- base_url: null
73
- endpoint_name: null
74
- endpoint_namespace: null
75
- generation_kwargs:
76
- max_new_tokens: 512
77
- temperature: 0.7
78
- model_display_name: null
79
- model_id: meta-llama/Meta-Llama-3-70B-Instruct
80
- structured_output: null
81
- tokenizer_id: meta-llama/Meta-Llama-3-70B-Instruct
82
- type_info:
83
- module: distilabel.llms.huggingface.inference_endpoints
84
- name: InferenceEndpointsLLM
85
- use_openai_client: false
86
- name: generate_sentence_pair
87
- num_generations: 1
88
- output_mappings:
89
- model_name: model_name_query
90
- runtime_parameters_info:
91
- - description: The number of rows that will contain the batches processed by
92
- the step.
93
- name: input_batch_size
94
- optional: true
95
- - name: llm
96
- runtime_parameters_info:
97
- - description: The kwargs to be propagated to either `generate` or `agenerate`
98
- methods within each `LLM`.
99
- keys:
100
- - description: the maximum number of new tokens that the model will generate. Defaults
101
- to `128`.
102
- name: max_new_tokens
103
- optional: true
104
- - description: the repetition penalty to use for the generation. Defaults to
105
- `0.0`. Only applies if `use_openai_client=True`.
106
- name: frequency_penalty
107
- optional: true
108
- - description: the presence penalty to use for the generation. Defaults
109
- to `0.0`. Only applies if `use_openai_client=True`.
110
- name: presence_penalty
111
- optional: true
112
- - description: the repetition penalty to use for the generation. Defaults to
113
- `None`. Only applies if `use_openai_client=False`.
114
- name: repetition_penalty
115
- optional: true
116
- - description: the temperature to use for the generation. Defaults to `1.0`.
117
- name: temperature
118
- optional: true
119
- - description: whether to use sampling for the generation. Defaults to `False`. Only
120
- applies if `use_openai_client=False`.
121
- name: do_sample
122
- optional: true
123
- - description: the top-k value to use for the generation. Defaults to `0.8`,
124
- since neither `0.0` nor `1.0` are valid values in TGI.
125
- name: top_k
126
- optional: true
127
- - description: the top-p value to use for the generation. Defaults to `1.0`.
128
- name: top_p
129
- optional: true
130
- - description: the typical-p value to use for the generation. Defaults to
131
- `0.5`.
132
- name: typical_p
133
- optional: true
134
- - description: either a single string or a list of strings containing the
135
- sequences to stop the generation at. Defaults to `None`, but will be
136
- set to the `tokenizer.eos_token` if available.
137
- name: stop_sequences
138
- optional: true
139
- - description: whether to return the full text of the completion or just
140
- the generated text. Defaults to `False`, meaning that only the generated
141
- text will be returned.
142
- name: return_full_text
143
- optional: true
144
- - description: the seed to use for the generation. Defaults to `None`.
145
- name: seed
146
- optional: true
147
- - description: whether to add the watermark to the generated text. Defaults
148
- to `None`.
149
- name: watermark
150
- optional: true
151
- name: generation_kwargs
152
- - description: The name of the Inference Endpoint to use for the LLM.
153
- name: endpoint_name
154
- optional: true
155
- - description: The namespace of the Inference Endpoint to use for the LLM.
156
- name: endpoint_namespace
157
- optional: true
158
- - description: The base URL to use for the Inference Endpoints API requests.
159
- name: base_url
160
- optional: true
161
- - description: The API key to authenticate the requests to the Inference Endpoints
162
- API.
163
- name: api_key
164
- optional: true
165
- - description: The structured output format to use across all the generations.
166
- name: structured_output
167
- optional: true
168
- - description: Whether to include the raw output of the LLM in the key `raw_output_<TASK_NAME>`
169
- of the `distilabel_metadata` dictionary output column
170
- name: add_raw_output
171
- optional: true
172
- - description: The number of generations to be produced per input.
173
- name: num_generations
174
- optional: true
175
- triplet: true
176
- type_info:
177
- module: distilabel.steps.tasks.sentence_transformers
178
- name: GenerateSentencePair
179
- - name: multiply_queries
180
- step:
181
- add_raw_output: true
182
- group_generations: false
183
- input_batch_size: 10
184
- input_mappings:
185
- query: positive
186
- llm:
187
- base_url: null
188
- endpoint_name: null
189
- endpoint_namespace: null
190
- generation_kwargs:
191
- max_new_tokens: 512
192
- temperature: 0.7
193
- model_display_name: null
194
- model_id: meta-llama/Meta-Llama-3-70B-Instruct
195
- structured_output: null
196
- tokenizer_id: meta-llama/Meta-Llama-3-70B-Instruct
197
- type_info:
198
- module: distilabel.llms.huggingface.inference_endpoints
199
- name: InferenceEndpointsLLM
200
- use_openai_client: false
201
- name: multiply_queries
202
- num_generations: 1
203
- num_queries: 3
204
- output_mappings:
205
- model_name: model_name_query_multiplied
206
- runtime_parameters_info:
207
- - description: The number of rows that will contain the batches processed by
208
- the step.
209
- name: input_batch_size
210
- optional: true
211
- - name: llm
212
- runtime_parameters_info:
213
- - description: The kwargs to be propagated to either `generate` or `agenerate`
214
- methods within each `LLM`.
215
- keys:
216
- - description: the maximum number of new tokens that the model will generate. Defaults
217
- to `128`.
218
- name: max_new_tokens
219
- optional: true
220
- - description: the repetition penalty to use for the generation. Defaults to
221
- `0.0`. Only applies if `use_openai_client=True`.
222
- name: frequency_penalty
223
- optional: true
224
- - description: the presence penalty to use for the generation. Defaults
225
- to `0.0`. Only applies if `use_openai_client=True`.
226
- name: presence_penalty
227
- optional: true
228
- - description: the repetition penalty to use for the generation. Defaults to
229
- `None`. Only applies if `use_openai_client=False`.
230
- name: repetition_penalty
231
- optional: true
232
- - description: the temperature to use for the generation. Defaults to `1.0`.
233
- name: temperature
234
- optional: true
235
- - description: whether to use sampling for the generation. Defaults to `False`. Only
236
- applies if `use_openai_client=False`.
237
- name: do_sample
238
- optional: true
239
- - description: the top-k value to use for the generation. Defaults to `0.8`,
240
- since neither `0.0` nor `1.0` are valid values in TGI.
241
- name: top_k
242
- optional: true
243
- - description: the top-p value to use for the generation. Defaults to `1.0`.
244
- name: top_p
245
- optional: true
246
- - description: the typical-p value to use for the generation. Defaults to
247
- `0.5`.
248
- name: typical_p
249
- optional: true
250
- - description: either a single string or a list of strings containing the
251
- sequences to stop the generation at. Defaults to `None`, but will be
252
- set to the `tokenizer.eos_token` if available.
253
- name: stop_sequences
254
- optional: true
255
- - description: whether to return the full text of the completion or just
256
- the generated text. Defaults to `False`, meaning that only the generated
257
- text will be returned.
258
- name: return_full_text
259
- optional: true
260
- - description: the seed to use for the generation. Defaults to `None`.
261
- name: seed
262
- optional: true
263
- - description: whether to add the watermark to the generated text. Defaults
264
- to `None`.
265
- name: watermark
266
- optional: true
267
- name: generation_kwargs
268
- - description: The name of the Inference Endpoint to use for the LLM.
269
- name: endpoint_name
270
- optional: true
271
- - description: The namespace of the Inference Endpoint to use for the LLM.
272
- name: endpoint_namespace
273
- optional: true
274
- - description: The base URL to use for the Inference Endpoints API requests.
275
- name: base_url
276
- optional: true
277
- - description: The API key to authenticate the requests to the Inference Endpoints
278
- API.
279
- name: api_key
280
- optional: true
281
- - description: The structured output format to use across all the generations.
282
- name: structured_output
283
- optional: true
284
- - description: Whether to include the raw output of the LLM in the key `raw_output_<TASK_NAME>`
285
- of the `distilabel_metadata` dictionary output column
286
- name: add_raw_output
287
- optional: true
288
- - description: The number of generations to be produced per input.
289
- name: num_generations
290
- optional: true
291
- system_prompt: You are an AI assistant helping to generate diverse examples.
292
- Ensure the generated queries are all in separated lines and preceded by a
293
- dash. Do not generate anything else or introduce the task.
294
- type_info:
295
- module: __main__
296
- name: MultipleQueries
297
- - name: merge_columns
298
- step:
299
- columns:
300
- '0': positive
301
- '1': queries
302
- input_batch_size: 50
303
- input_mappings: {}
304
- name: merge_columns
305
- output_columns:
306
- '0': positive
307
- output_mappings: {}
308
- runtime_parameters_info:
309
- - description: The number of rows that will contain the batches processed by
310
- the step.
311
- name: input_batch_size
312
- optional: true
313
- type_info:
314
- module: distilabel.steps.combine
315
- name: CombineColumns
316
- - name: expand_columns_0
317
- step:
318
- columns:
319
- positive: positive
320
- input_batch_size: 50
321
- input_mappings: {}
322
- name: expand_columns_0
323
- output_mappings: {}
324
- runtime_parameters_info:
325
- - description: The number of rows that will contain the batches processed by
326
- the step.
327
- name: input_batch_size
328
- optional: true
329
- type_info:
330
- module: distilabel.steps.expand
331
- name: ExpandColumns
332
- type_info:
333
- module: distilabel.pipeline.local
334
- name: Pipeline