From 73e661e4dbdaeac2b158c2a9a89ec5cd05626161 Mon Sep 17 00:00:00 2001
From: Julien Breton <julien.breton@moncitron.fr>
Date: Sun, 4 Feb 2024 14:49:23 +0900
Subject: [PATCH] change hyper parameters
---
modules/llm/Mistral-7b/Mistral-7b_fine_tune.py | 1 -
modules/llm/Mixtral-8x7b/Mixtral-8x7b_fine_tune.py | 1 -
2 files changed, 2 deletions(-)
diff --git a/modules/llm/Mistral-7b/Mistral-7b_fine_tune.py b/modules/llm/Mistral-7b/Mistral-7b_fine_tune.py
index a41fc61..99a152d 100644
--- a/modules/llm/Mistral-7b/Mistral-7b_fine_tune.py
+++ b/modules/llm/Mistral-7b/Mistral-7b_fine_tune.py
@@ -58,7 +58,6 @@ def fine_tune(base_model, new_model):
training_arguments = TrainingArguments(
per_device_train_batch_size=2,
gradient_accumulation_steps=1,
- num_train_epochs=2,
learning_rate=1e-4,
logging_steps=2,
optim="adamw_torch",
diff --git a/modules/llm/Mixtral-8x7b/Mixtral-8x7b_fine_tune.py b/modules/llm/Mixtral-8x7b/Mixtral-8x7b_fine_tune.py
index d9e8126..9210d02 100644
--- a/modules/llm/Mixtral-8x7b/Mixtral-8x7b_fine_tune.py
+++ b/modules/llm/Mixtral-8x7b/Mixtral-8x7b_fine_tune.py
@@ -72,7 +72,6 @@ def fine_tuned(base_model, new_model):
args=TrainingArguments(
per_device_train_batch_size=2,
gradient_accumulation_steps=1,
- num_train_epochs=2,
learning_rate=1e-4,
logging_steps=2,
optim="adamw_torch",
--
GitLab