v3.2 hotfix with some corrections

.gitignore

@@ -3,6 +3,8 @@
 /src/data/syriac**.jsonl
 /src/data/sedra**.jsonl
 /src/data/corpus**.jsonl
+/src/data/correction**.jsonl
+/src/finetune_corrections.py
 /src/data/**.txt
 /logs/**
 /checkpoints/**

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b0dc22e4241c40835061a8408785a0b93cf5b139725732e806e28115dda2b26d
+oid sha256:366187761c180de342bc1d86daa7191034d2b0e70a34d7724081c4f9bde8d638
 size 209216552

src/test_t5.py

@@ -1,7 +1,7 @@
 from transformers import AutoTokenizer, T5ForConditionalGeneration, pipeline
 
 # HF Hub path config
-model_path = "crossroderick/aramt5"
+model_path = "checkpoints/aramt5"
 
 # Unicode directional formatting for RTL text (Syriac)
 RLI = "\u2067"  # Right-to-Left Isolate

tokenizer.json

@@ -1,6 +1,11 @@
 {
   "version": "1.0",
-  "truncation": null,
+  "truncation": {
+    "direction": "Right",
+    "max_length": 128,
+    "strategy": "LongestFirst",
+    "stride": 0
+  },
   "padding": null,
   "added_tokens": [
     {

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57859de1703fd61d7e4a8046b1f9eb356e9e8313bf4da88e5f43f4118c24b191
+size 5905