Guide-yakabatana mibvunzo · Pakati Nhanho
Kuenderera mberi nePretraining vs Fine-Tuning Quiz
Inoedza dhata uye misiyano yechinangwa pakati pekuenderera mberi kwekudzidzira nekugadzirisa zvakanaka, kukanganwa kunotyisa nekudzokorora, uye kana kudzoreredza kuchikwana zvirinani.
Mubvunzo 1 ye 8
Chii chisingafanirwe pakati pekudzidzisa kwekufambidzira nekudzidzisa kwekugadzirisa?
Tarisa mubvunzo wese8 mibvunzo
Mhinduro uye tsananguro zvinoonekwa paunenge uchitora bvunzo.
- Chii chisingafanirwe pakati pekudzidzisa kwekufambidzira nekudzidzisa kwekugadzirisa?
- Both use only labeled question and answer pairs
- Continued pretraining uses raw domain text; fine-tuning uses curated labeled examples
- Fine-tuning uses more data than continued pretraining
- Continued pretraining uses only images
- Mune kudzidzisa kwekugadzirisa, panzi kusawirirana kunowanzoitwa?
- On every token including the prompt
- Only on the prompt tokens
- On randomly chosen tokens
- Only on the response tokens
- Replay ndeipi mune kudzidzisa kwekufambidzira?
- Mixing some general data into training to limit forgetting
- Running the same batch twice
- Restarting training from scratch
- Replaying user conversations at inference time
- Code Llama yakagadziriswa sei?
- From scratch on code only
- By prompt engineering Llama 2
- By continuing to train Llama 2 on a large code corpus
- By fine-tuning on a few hundred coding questions
- Kune zvinangwa zvinosiyana-siyana, nzira ipi inotsanangura kuti inokosha kupfuura kudzidzisa kwekugadzirisa?
- Continued pretraining every day
- Retrieval of up-to-date documents
- Raising the sampling temperature
- Training a larger tokenizer
- Which statement about BloombergGPT is correct according to the guide?
- It was continued pretraining of Llama 2
- It was created only with LoRA
- It was a prompt-engineered GPT-3
- It was trained from scratch on financial and general data
- What did the 2020 paper Don't Stop Pretraining show?
- A second phase of pretraining on in-domain text improved RoBERTa on domain tasks
- Pretraining should always stop after one epoch
- Fine-tuning is useless for classification
- Domain text harms every model
- Why re-warm the learning rate at the start of continued pretraining?
- To reset the tokenizer
- To freeze the base weights
- Resuming at a high rate destabilizes training while too low a rate learns little
- To skip the need for domain data