best model, 7K steps

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "facebook/m2m100_1.2B",
+  "_name_or_path": "anzorq/m2m100_1.2B_ft_ru-kbd_50K",
   "activation_dropout": 0.0,
   "activation_function": "relu",
   "architectures": [
@@ -32,7 +32,7 @@
   "pad_token_id": 1,
   "scale_embedding": true,
   "torch_dtype": "float32",
-  "transformers_version": "4.34.0.dev0",
+  "transformers_version": "4.33.3",
   "use_cache": true,
   "vocab_size": 128112
 }

generation_config.json

@@ -0,0 +1,12 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 0,
+  "decoder_start_token_id": 2,
+  "early_stopping": true,
+  "eos_token_id": 2,
+  "forced_bos_token_id": 128103,
+  "max_length": 200,
+  "num_beams": 5,
+  "pad_token_id": 1,
+  "transformers_version": "4.33.3"
+}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ccce11adab46b8d2f1df26383529a31d908d8c2a2e791cefa770c2e65bab27e7
+oid sha256:9dd1b38cc3e6ed06b175f422110121988443f701ca8ccb3ea3e6cb0f66909038
 size 4958228362