improve KTO impl., replace datasets

Former-commit-id: e56a57ddcf061de6e4acc8679f7dbf0b68364986
2024-05-18 03:44:56 +08:00
parent e4570e28a8
commit 2bff90719b
53 changed files with 448 additions and 330 deletions
--- a/examples/qlora_single_gpu/llama3_lora_sft_aqlm.yaml
+++ b/examples/qlora_single_gpu/llama3_lora_sft_aqlm.yaml
@@ -8,7 +8,7 @@ finetuning_type: lora
 lora_target: q_proj,v_proj

 ### dataset
-dataset: identity,alpaca_gpt4_en
+dataset: identity,alpaca_en_demo
 template: llama3
 cutoff_len: 1024
 max_samples: 1000
--- a/examples/qlora_single_gpu/llama3_lora_sft_awq.yaml
+++ b/examples/qlora_single_gpu/llama3_lora_sft_awq.yaml
@@ -8,7 +8,7 @@ finetuning_type: lora
 lora_target: q_proj,v_proj

 ### dataset
-dataset: identity,alpaca_gpt4_en
+dataset: identity,alpaca_en_demo
 template: llama3
 cutoff_len: 1024
 max_samples: 1000
--- a/examples/qlora_single_gpu/llama3_lora_sft_bitsandbytes.yaml
+++ b/examples/qlora_single_gpu/llama3_lora_sft_bitsandbytes.yaml
@@ -9,7 +9,7 @@ finetuning_type: lora
 lora_target: q_proj,v_proj

 ### dataset
-dataset: identity,alpaca_gpt4_en
+dataset: identity,alpaca_en_demo
 template: llama3
 cutoff_len: 1024
 max_samples: 1000
--- a/examples/qlora_single_gpu/llama3_lora_sft_gptq.yaml
+++ b/examples/qlora_single_gpu/llama3_lora_sft_gptq.yaml
@@ -8,7 +8,7 @@ finetuning_type: lora
 lora_target: q_proj,v_proj

 ### dataset
-dataset: identity,alpaca_gpt4_en
+dataset: identity,alpaca_en_demo
 template: llama3
 cutoff_len: 1024
 max_samples: 1000