| <!--Copyright 2020 The HuggingFace Team. All rights reserved. | |
| Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with | |
| the License. You may obtain a copy of the License at | |
| http://www.apache.org/licenses/LICENSE-2.0 | |
| Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on | |
| an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the | |
| specific language governing permissions and limitations under the License. | |
| β οΈ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be | |
| rendered properly in your Markdown viewer. | |
| --> | |
| # Trainer [[trainer]] | |
| [`Trainer`] ν΄λμ€λ PyTorchμμ μμ ν κΈ°λ₯(feature-complete)μ νλ ¨μ μν APIλ₯Ό μ 곡νλ©°, λ€μ€ GPU/TPUμμμ λΆμ° νλ ¨, [NVIDIA GPU](https://nvidia.github.io/apex/), [AMD GPU](https://rocm.docs.amd.com/en/latest/rocm.html)λ₯Ό μν νΌν© μ λ°λ, κ·Έλ¦¬κ³ PyTorchμ [`torch.amp`](https://pytorch.org/docs/stable/amp.html)λ₯Ό μ§μν©λλ€. [`Trainer`]λ λͺ¨λΈμ νλ ¨ λ°©μμ 컀μ€ν°λ§μ΄μ¦ν μ μλ λ€μν μ΅μ μ μ 곡νλ [`TrainingArguments`] ν΄λμ€μ ν¨κ» μ¬μ©λ©λλ€. μ΄ λ ν΄λμ€λ ν¨κ» μμ ν νλ ¨ APIλ₯Ό μ 곡ν©λλ€. | |
| [`Seq2SeqTrainer`]μ [`Seq2SeqTrainingArguments`]λ [`Trainer`]μ [`TrainingArguments`] ν΄λμ€λ₯Ό μμνλ©°, μμ½μ΄λ λ²μκ³Ό κ°μ μνμ€-ν¬-μνμ€ μμ μ μν λͺ¨λΈ νλ ¨μ μ ν©νκ² μ‘°μ λμ΄ μμ΅λλ€. | |
| <Tip warning={true}> | |
| [`Trainer`] ν΄λμ€λ π€ Transformers λͺ¨λΈμ μ΅μ νλμ΄ μμΌλ©°, λ€λ₯Έ λͺ¨λΈκ³Ό ν¨κ» μ¬μ©λ λ μμμΉ λͺ»ν λμμ νκ² λ μ μμ΅λλ€. μμ λ§μ λͺ¨λΈμ μ¬μ©ν λλ λ€μμ νμΈνμΈμ: | |
| - λͺ¨λΈμ νμ ννμ΄λ [`~utils.ModelOutput`]μ μλΈν΄λμ€λ₯Ό λ°νν΄μΌ ν©λλ€. | |
| - λͺ¨λΈμ `labels` μΈμκ° μ 곡λλ©΄ μμ€μ κ³μ°ν μ μκ³ , λͺ¨λΈμ΄ ννμ λ°ννλ κ²½μ° κ·Έ μμ€μ΄ ννμ 첫 λ²μ§Έ μμλ‘ λ°νλμ΄μΌ ν©λλ€. | |
| - λͺ¨λΈμ μ¬λ¬ κ°μ λ μ΄λΈ μΈμλ₯Ό μμ©ν μ μμ΄μΌ νλ©°, [`Trainer`]μκ² μ΄λ¦μ μ리기 μν΄ [`TrainingArguments`]μμ `label_names`λ₯Ό μ¬μ©νμ§λ§, κ·Έ μ€ μ΄λ κ²λ `"label"`λ‘ λͺ λͺ λμ΄μλ μ λ©λλ€. | |
| </Tip> | |
| ## Trainer [[transformers.Trainer]] | |
| [[autodoc]] Trainer | |
| - all | |
| ## Seq2SeqTrainer [[transformers.Seq2SeqTrainer]] | |
| [[autodoc]] Seq2SeqTrainer | |
| - evaluate | |
| - predict | |
| ## TrainingArguments [[transformers.TrainingArguments]] | |
| [[autodoc]] TrainingArguments | |
| - all | |
| ## Seq2SeqTrainingArguments [[transformers.Seq2SeqTrainingArguments]] | |
| [[autodoc]] Seq2SeqTrainingArguments | |
| - all | |