From b731e729d2dd61376c3c696476f13b7198bb24ee Mon Sep 17 00:00:00 2001
From: Abhishek Thakur <abhishek4@gmail.com>
Date: Wed, 14 Aug 2024 14:21:43 +0200
Subject: [PATCH] add seq2seq configs

---
 configs/seq2seq/hub_dataset.yml | 28 ++++++++++++++++++++++++++++
 configs/seq2seq/local.yml       | 29 +++++++++++++++++++++++++++++
 2 files changed, 57 insertions(+)
 create mode 100644 configs/seq2seq/hub_dataset.yml
 create mode 100644 configs/seq2seq/local.yml

diff --git a/configs/seq2seq/hub_dataset.yml b/configs/seq2seq/hub_dataset.yml
new file mode 100644
index 0000000000..5df712162b
--- /dev/null
+++ b/configs/seq2seq/hub_dataset.yml
@@ -0,0 +1,28 @@
+task: seq2seq
+base_model: google/flan-t5-base
+project_name: autotrain-seq2seq-hub-dataset
+log: tensorboard
+backend: local
+
+data:
+  path: samsum
+  train_split: train
+  valid_split: test
+  column_mapping:
+    text_column: dialogue
+    target_column: summary
+
+params:
+  max_seq_length: 512
+  epochs: 3
+  batch_size: 4
+  lr: 2e-5
+  optimizer: adamw_torch
+  scheduler: linear
+  gradient_accumulation: 1
+  mixed_precision: none
+
+hub:
+  username: ${HF_USERNAME}
+  token: ${HF_TOKEN}
+  push_to_hub: true
\ No newline at end of file
diff --git a/configs/seq2seq/local.yml b/configs/seq2seq/local.yml
new file mode 100644
index 0000000000..16ba8ad69f
--- /dev/null
+++ b/configs/seq2seq/local.yml
@@ -0,0 +1,29 @@
+task: seq2seq
+base_model: google/flan-t5-base
+project_name: autotrain-seq2seq-local
+log: tensorboard
+backend: local
+
+data:
+  path: path/to/your/dataset csv/jsonl files
+  train_split: train
+  valid_split: test
+  column_mapping:
+    text_column: text
+    target_column: target
+
+
+params:
+  max_seq_length: 512
+  epochs: 3
+  batch_size: 4
+  lr: 2e-5
+  optimizer: adamw_torch
+  scheduler: linear
+  gradient_accumulation: 1
+  mixed_precision: none
+
+hub:
+  username: ${HF_USERNAME}
+  token: ${HF_TOKEN}
+  push_to_hub: true
\ No newline at end of file