Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Skill:List specific number synonyms for one word (#1130)
If your PR is related to a contribution to the taxonomy, please, fill out the following questionnaire. If not, replace this whole text and the following questionnaire with whatever information is applicable to your PR. **Describe the contribution to the taxonomy** <!-- A concise description of what the contribution brings, replace "..." in the bullet list --> - ... List specific number synonyms for one word - ... List one synonyms for one word - ...List two synonyms for one word - ... List three synonyms for one word **Input given at the prompt** <!-- What you entered, replace "..." --> ``` (instructlab_vv) [root@dell-r640-041 synonyms]# ilab generate --num-instructions 10 llama_cpp_python is built without hardware acceleration. ilab generate will be very slow. Generating synthetic data using 'merlinite-7b-lab-Q4_K_M' model, taxonomy:'/home/instructlab/taxonomy' against http://127.0.0.1:8000/v1 server INFO 2024-05-26 22:36:58,186 generate_data.py:468 Selected taxonomy path compositional_skills->general->synonyms Q> What is an antonym to the word ‘attend’? Please list and separate with newline. I> A> An antonym to Attend is not attend. 10%|████████▌ | 1/10 [00:20<03:05, 20.58s/it]INFO 2024-05-26 22:37:18,766 generate_data.py:468 Selected taxonomy path compositional_skills->general->synonyms INFO 2024-05-26 22:37:38,998 generate_data.py:468 Selected taxonomy path compositional_skills->general->synonyms INFO 2024-05-26 22:37:44,410 generate_data.py:468 Selected taxonomy path compositional_skills->general->synonyms INFO 2024-05-26 22:38:42,586 generate_data.py:468 Selected taxonomy path compositional_skills->general->synonyms INFO 2024-05-26 22:38:55,369 generate_data.py:468 Selected taxonomy path compositional_skills->general->synonyms Q> Please list and separate with newline some alternative words to describe the word 'attend' that are used when a person is participating in an event or a gathering. I> A> Alternative words to describe 'Attend' are: Participate in Take part in Show up at Be present at Join in Take part in Contribute to Make an appearance at Turn up for Come to 30%|█████████████████████████▌ | 3/10 [02:33<06:22, 50%|██████████████████████████████████████████▌ | 60%|███████████████████████████████████████████████████ 90%|████████████████████████████████████████████████████████████████████████████▌ | 9/10 [07:32<00:48, 48.13s/it]INFO 2024-05-26 22:44:30,664 generate_data.py:468 Selected taxonomy path compositional_skills->general->synonyms INFO 2024-05-26 22:44:52,830 generate_data.py:468 Selected taxonomy path compositional_skills->general->synonyms INFO 2024-05-26 22:45:40,164 generate_data.py:468 Selected taxonomy path compositional_skills->general->synonyms INFO 2024-05-26 22:45:51,975 generate_data.py:468 Selected taxonomy path compositional_skills->general->synonyms INFO 2024-05-26 22:45:58,184 generate_data.py:468 Selected taxonomy path compositional_skills->general->synonyms INFO 2024-05-26 22:46:21,326 generate_data.py:468 Selected taxonomy path compositional_skills->general->synonyms INFO 2024-05-26 22:46:46,113 generate_data.py:468 Selected taxonomy path compositional_skills->general->synonyms ``` ``` (instructlab_vv) [root@dell-r640-041 synonyms]# ilab train INFO 2024-05-26 22:48:53,726 config.py:58 PyTorch version 2.3.0 available. LINUX_TRAIN.PY: NUM EPOCHS IS: 1 LINUX_TRAIN.PY: TRAIN FILE IS: generated/train_merlinite-7b-lab-Q4_K_M_2024-05-26T22_36_58.jsonl LINUX_TRAIN.PY: TEST FILE IS: generated/test_merlinite-7b-lab-Q4_K_M_2024-05-26T22_36_58.jsonl LINUX_TRAIN.PY: Using device 'cpu' LINUX_TRAIN.PY: LOADING DATASETS Generating train split: 11 examples [00:00, 3144.80 examples/s] Generating train split: 4 examples [00:00, 1685.64 examples/s] /home/env/instructlabenv/instructlab_vv/lib/python3.12/site-packages/huggingface_hub/file_download.py:1132: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`. warnings.warn( tokenizer_config.json: 100%|████████████████████████████████████████████████████████| 2.33k/2.33k [00:00<00:00, 23.2MB/s] tokenizer.model: 100%|████████████████████████████████████████████████████████████████| 493k/493k [00:00<00:00, 9.62MB/s] tokenizer.json: 100%|███████████████████████████████████████████████████████████████| 1.80M/1.80M [00:00<00:00, 22.8MB/s] added_tokens.json: 100%|████████████████████████████████████████████████████████████████| 119/119 [00:00<00:00, 1.52MB/s] special_tokens_map.json: 100%|██████████████████████████████████████████████████████████| 655/655 [00:00<00:00, 7.99MB/s] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. LINUX_TRAIN.PY: NOT USING 4-bit quantization LINUX_TRAIN.PY: LOADING THE BASE MODEL config.json: 100%|██████████████████████████████████████████████████████████████████████| 644/644 [00:00<00:00, 6.84MB/s] model.safetensors.index.json: 100%|██████████████████████████████████████████████████| 23.9k/23.9k [00:00<00:00, 161MB/s] model-00001-of-00003.safetensors: 100%|██████████████████████████████████████████████| 4.94G/4.94G [00:43<00:00, 115MB/s] model-00002-of-00003.safetensors: 100%|██████████████████████████████████████████████| 5.00G/5.00G [00:43<00:00, 115MB/s] model-00003-of-00003.safetensors: 100%|██████████████████████████████████████████████| 4.54G/4.54G [00:39<00:00, 115MB/s] Downloading shards: 100%|██████████████████████████████████████████████████████████████████| 3/3 [02:06<00:00, 42.12s/it] Loading checkpoint shards: 100%|███████████████████████████████████████████████████████████| 3/3 [00:00<00:00, 6.19it/s] generation_config.json: 100%|███████████████████████████████████████████████████████████| 136/136 [00:00<00:00, 1.53MB/s] LINUX_TRAIN.PY: Model device cpu LINUX_TRAIN.PY: SANITY CHECKING THE BASE MODEL 100%|██████████████████████████████████████████████████████████████████████████████████████| 4/4 [02:40<00:00, 40.04s/it] LINUX_TRAIN.PY: GETTING THE ATTENTION LAYERS LINUX_TRAIN.PY: CONFIGURING LoRA Map: 100%|██████████████████████████████████████████████████████████████████████| 11/11 [00:00<00:00, 1414.65 examples/s] Map: 100%|████████████████████████████████████████████████████████████████████████| 4/4 [00:00<00:00, 1006.37 examples/s] /home/env/instructlabenv/instructlab_vv/lib/python3.12/site-packages/accelerate/accelerator.py:446: FutureWarning: Passing the following arguments to `Accelerator` is deprecated and will be removed in version 1.0 of Accelerate: dict_keys(['dispatch_batches', 'split_batches', 'even_batches', 'use_seedable_sampler']). Please pass an `accelerate.DataLoaderConfiguration` instead: dataloader_config = DataLoaderConfiguration(dispatch_batches=None, split_batches=False, even_batches=True, use_seedable_sampler=True) warnings.warn( LINUX_TRAIN.PY: TRAINING {'train_runtime': 207.9129, 'train_samples_per_second': 0.053, 'train_steps_per_second': 0.053, 'train_loss': 1.4257791692560369, 'epoch': 1.0} 100%|████████████████████████████████████████████████████████████████████████████████████| 11/11 [03:27<00:00, 18.90s/it] LINUX_TRAIN.PY: RUNNING INFERENCE ON THE OUTPUT MODEL === test 0 === === user === List a synonym for the word attend. === assistant_old === Certainly! A synonym for the word "attend" is "participate." Both words have a similar meaning, indicating involvement or engagement in an event or activity. For example, you can "attend" a class or meeting, or you can "participate" in a discussion or workshop. === assistant_new === Attend: The word attend has several synonyms, including: - Be present - Show up - Be there - Participate - Be in attendance - Be present at - Show up for - Be a part of - Be involved in - Join - Be there for - Be present with - Be on hand - Be on location - Be in the audience - Be in the crowd - Be in the congregation - Be in the gathering - Be in the assembly - Be in the meeting - Be in the session - Be in the class - Be in the lecture - Be in the seminar - Be in the workshop - Be in the conference - Be in the symposium - Be in the forum - Be in the discussion - Be in the debate - Be in the interview - Be in the talk - Be in the speech - Be in the presentation - Be in the performance - Be in the show - Be === assistant_expected === Synonym for Attend is take part in === test 1 === === user === List two synonyms for the word attend and separate with newline. === assistant_old === attend attend attend attend attend attend attend attend attend attend attend attend attend attend attend attend attend attend attend === assistant_new === Attend\n Participate\n === assistant_expected === Two synonyms for Attend are take part in be present at === test 2 === === user === List three synonyms for the word attend and separate with newline. === assistant_old === Attend, attend to, and attend upon are three synonyms for the word attend. Attend: To be present, to be in attendance. Attend to: To give attention to, to take care of, to see to. Attend upon: To wait upon, to serve, to minister to. These synonyms can be used interchangeably in various contexts to convey the meaning of being present or giving attention to something or someone. For example: * I will attend the meeting at 3 PM. * Please attend to the customer's needs. * I will attend upon the king during his visit. These examples illustrate how the synonyms can be used in sentences to convey the meaning of being present or giving attention. === assistant_new === 1. attend 2. attend 3. attend === assistant_expected === Three synonyms for Attend are take part in be present at be there at === test 3 === === user === List four synonyms for the word attend and separate with newline. === assistant_old === 1. attend 2. attend 3. attend 4. attend I'm sorry for the mistake in my previous comment. I understand that you were asking for synonyms for the word "attend." However, I misunderstood your request. Here are four synonyms for "attend": 1. assist 2. accompany 3. participate 4. observe I apologize for any confusion and hope this information is helpful. If you have any other questions, please let me know! === assistant_new === 1. attend \n 2. be present \n 3. be present at \n 4. be present in \n === assistant_expected === Four synonyms for Attend are take part in be present at be there at make an appearance at Wrote training_results/final/ggml-model-f16.gguf ``` **Response from the fine-tuned model** <!-- Generate a synthetic dataset based on your newly added seed data; train the model with the synthetic data and now re-test the model's response with the same prompt. Replace "..." with what you receive with the finetuned model. --> ``` (instructlab_vv) [root@dell-r640-041 synonyms]# ilab chat ╭─────────────────────────────────────────────────────── system ────────────────────────────────────────────────────────╮ │ Welcome to InstructLab Chat w/ MODELS/MERLINITE-7B-LAB-Q4_K_M.GGUF (type /h for help) │ ╰─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────── elapsed 5.047 seconds ─╯ >>> List a synonym for the word attend. [S][default] ╭───────────────────────────────────────── models/merlinite-7b-lab-Q4_K_M.gguf ─────────────────────────────────────────╮ │ A synonym for the word "attend" is "show up." │ │ │ │ (continued from previous session) │ ╰─────────────────────────────────────────────────────────────────────────────────────────────── elapsed 3.311 seconds ─╯ >>> List two synonyms for the word attend [S][default] ╭───────────────────────────────────────── models/merlinite-7b-lab-Q4_K_M.gguf ─────────────────────────────────────────╮ │ Two synonyms for the word "attend" are "present yourself" and "make an appearance." │ │ │ │ (continued from previous session) │ ╰─────────────────────────────────────────────────────────────────────────────────────────────── elapsed 4.259 seconds ─╯ >>> List two synonyms for the word beat [S][default] ╭───────────────────────────────────────── models/merlinite-7b-lab-Q4_K_M.gguf ─────────────────────────────────────────╮ │ Two synonyms for the word "beat" are "thump" and "drumbeat." │ │ │ │ (continued from previous session) │ ╰──────────────────────────────────────────────────``` **Contribution checklist** <!-- Insert an x between the empty brackets: [ ] >> [x] --> - [ ] The contribution was tested with `ilab generate` - [ ] No errors or warnings were produced by `ilab generate` - [ ] All [commits are signed off](https://github.com/instructlab/taxonomy/blob/main/CONTRIBUTING.md#legal) (DCO) - [ ] The `qna.yaml` file contains at least 5 `seed_examples` - [ ] The `qna.yaml` file was [linted](https://yamllint.com) and [prettified](https://onlineyamltools.com/prettify-yaml) ([yaml-validator](https://jsonformatter.org/yaml-validator) can do both) - [ ] An `attribution.txt` file in the same folder as the `qna.yaml` file - [ ] Content does not include PII or otherwise sensitive or confidential information - [ ] Content does not include anything documented in the project's [Avoid these Topics](https://github.com/instructlab/taxonomy/blob/main/docs/SKILLS_GUIDE.md#avoid-these-topics) guidelines --------- Signed-off-by: yanpliu <[email protected]> Signed-off-by: JJ Asghar <[email protected]> Co-authored-by: JJ Asghar <[email protected]>
- Loading branch information
