added hf configs

Helsinki-NLP · Sep 20, 2024 · 9f60d34 · 9f60d34
1 parent 221451c
commit 9f60d34
Show file tree

Hide file tree

Showing 5 changed files with 12 additions and 29 deletions.
diff --git a/configs/hf/config.hf.hplt.yml b/configs/hf/config.hf.hplt.yml
@@ -5,21 +5,17 @@
 ###
 
 experiment:
-  dirname: test
-  name: hf
+  dirname: hf
+  name: hplt
   langpairs:
     - et-en
-  split-length: 10000
+
+  split-length: 1000
   best-model: perplexity
 
   huggingface:
     modelname: "HPLT/translate-et-en-v1.0-hplt_opus"
     modelclass: "transformers.AutoModelForSeq2SeqLM"
-    config:
-      top_k: 50
-      top_p: 0.90
-      temperature: 0.1
-      max_new_tokens: 128
 
 datasets:
   train:

diff --git a/configs/hf/config.hf.nllb.yml b/configs/hf/config.hf.nllb.yml
@@ -8,13 +8,9 @@ experiment:
   dirname: hf
   name: nllb
   langpairs:
-    - en-fi
     - en-et
 
-  parallel-max-sentences: 1000
-  split-length: 10000
-  spm-sample-size: 400
-  spm-vocab-size: 3000
+  split-length: 1000
 
   best-model: perplexity
 

diff --git a/configs/hf/config.hf.opus.yml b/configs/hf/config.hf.opus.yml
@@ -8,13 +8,9 @@ experiment:
   dirname: hf
   name: opus
   langpairs:
-    - en-fi
     - en-et
 
-  parallel-max-sentences: 1000
-  split-length: 10000
-  spm-sample-size: 400
-  spm-vocab-size: 3000
+  split-length: 1000
 
   best-model: perplexity
 

diff --git a/configs/hf/config.hf.t5.yml b/configs/hf/config.hf.t5.yml
@@ -10,10 +10,7 @@ experiment:
   langpairs:
     - en-de
 
-  parallel-max-sentences: 1000
-  split-length: 10000
-  spm-sample-size: 400
-  spm-vocab-size: 3000
+  split-length: 1000
 
   best-model: perplexity
 
@@ -24,10 +21,12 @@ experiment:
       en: English
       de: German
     prompt: "Translate {src_lang} to {tgt_lang}: {source}"
+    config:
+      max_new_tokens: 150
 
 datasets:
   train:
-    - ELRC-3293-EUROPARL_covid__v1
+    - ELRC_2922__v1
   devtest:
     - flores_dev
   test:

diff --git a/configs/hf/config.utter.yml b/configs/hf/config.utter.yml
@@ -8,14 +8,10 @@ experiment:
   dirname: hf
   name: euroLLM
   langpairs:
-    - en-fi
     - en-et
 
-  parallel-max-sentences: 1000
-  split-length: 10000
-  spm-sample-size: 400
-  spm-vocab-size: 3000
-
+  split-length: 1000
+
   best-model: perplexity
 
   huggingface: