migrate example configs to canonical attn_implementation

2026-04-23 22:15:07 +00:00
parent 2d64d009d8
commit 39226623d2
222 changed files with 209 additions and 243 deletions
--- a/examples/qat_nvfp4/Gemma3-12B_baseline.yml
+++ b/examples/qat_nvfp4/Gemma3-12B_baseline.yml
@@ -24,7 +24,7 @@ output_dir: ./outputs/out_gemma/

 sequence_len: 8096
 sample_packing: true
-flash_attention: true
+attn_implementation: flash_attention_2

 wandb_entity:
 wandb_watch:
--- a/examples/qat_nvfp4/Gemma3-12B_qat.yml
+++ b/examples/qat_nvfp4/Gemma3-12B_qat.yml
@@ -24,7 +24,7 @@ output_dir: ./outputs/qat_out_gemma/

 sequence_len: 8096
 sample_packing: true
-flash_attention: true
+attn_implementation: flash_attention_2

 qat:
  activation_dtype: nvfp4
--- a/examples/qat_nvfp4/Math-Gemma3-12B_baseline.yml
+++ b/examples/qat_nvfp4/Math-Gemma3-12B_baseline.yml
@@ -24,7 +24,7 @@ output_dir: ./outputs/out_math_gemma/

 sequence_len: 4096
 sample_packing: true
-flash_attention: true
+attn_implementation: flash_attention_2

 wandb_entity:
 wandb_watch:
--- a/examples/qat_nvfp4/Math-Gemma3-12B_qat.yml
+++ b/examples/qat_nvfp4/Math-Gemma3-12B_qat.yml
@@ -24,7 +24,7 @@ output_dir: ./outputs/qat_out_math_gemma/

 sequence_len: 4096
 sample_packing: true
-flash_attention: true
+attn_implementation: flash_attention_2

 qat:
  activation_dtype: nvfp4
--- a/examples/qat_nvfp4/Math-Gemma3-27B_baseline.yml
+++ b/examples/qat_nvfp4/Math-Gemma3-27B_baseline.yml
@@ -24,7 +24,7 @@ output_dir: ./outputs/out_math_gemma27/

 sequence_len: 4096
 sample_packing: true
-flash_attention: true
+attn_implementation: flash_attention_2

 wandb_entity:
 wandb_watch:
--- a/examples/qat_nvfp4/Math-Gemma3-27B_qat.yml
+++ b/examples/qat_nvfp4/Math-Gemma3-27B_qat.yml
@@ -24,7 +24,7 @@ output_dir: ./outputs/qat_out_math_gemma27/

 sequence_len: 4096
 sample_packing: true
-flash_attention: true
+attn_implementation: flash_attention_2

 qat:
  activation_dtype: nvfp4
--- a/examples/qat_nvfp4/Math-Qwen2.5-72B_baseline.yml
+++ b/examples/qat_nvfp4/Math-Qwen2.5-72B_baseline.yml
@@ -24,7 +24,7 @@ output_dir: ./outputs/out_math_72b/

 sequence_len: 4096
 sample_packing: true
-flash_attention: true
+attn_implementation: flash_attention_2

 wandb_entity:
 wandb_watch:
--- a/examples/qat_nvfp4/Math-Qwen2.5-72B_qat.yml
+++ b/examples/qat_nvfp4/Math-Qwen2.5-72B_qat.yml
@@ -24,7 +24,7 @@ output_dir: ./outputs/qat_out_math_72b/

 sequence_len: 4096
 sample_packing: true
-flash_attention: true
+attn_implementation: flash_attention_2

 qat:
  activation_dtype: nvfp4
--- a/examples/qat_nvfp4/Qwen2.5-72B_baseline.yml
+++ b/examples/qat_nvfp4/Qwen2.5-72B_baseline.yml
@@ -24,7 +24,7 @@ output_dir: ./outputs/out_qwen72b/

 sequence_len: 8096
 sample_packing: true
-flash_attention: true
+attn_implementation: flash_attention_2

 wandb_entity:
 wandb_watch:
--- a/examples/qat_nvfp4/Qwen2.5-72B_qat.yml
+++ b/examples/qat_nvfp4/Qwen2.5-72B_qat.yml
@@ -24,7 +24,7 @@ output_dir: ./outputs/qat_out_qwen72b/

 sequence_len: 8096
 sample_packing: true
-flash_attention: true
+attn_implementation: flash_attention_2

 qat:
  activation_dtype: nvfp4