Training in progress, step 10

adapter_config.json

@@ -1,6 +1,6 @@
 {
   "auto_mapping": null,
-  "base_model_name_or_path": "tiiuae/falcon-rw-1b",
+  "base_model_name_or_path": "tiiuae/falcon-7b",
   "bias": "none",
   "fan_in_fan_out": false,
   "inference_mode": true,
@@ -11,7 +11,7 @@
   "lora_dropout": 0.05,
   "modules_to_save": null,
   "peft_type": "LORA",
-  "r": 16,
+  "r": 32,
   "revision": null,
   "target_modules": [
     "query_key_value",

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b92153cb621b0e67b71df37db519893470ee29e08635b054d1c4b3a07e0232fc
-size 50401805
+oid sha256:b86aa3744b927b6290d89922f5d38371b27acef777d25b41d81278bc4851f427
+size 261189453

special_tokens_map.json

@@ -1,6 +1,17 @@
 {
-  "bos_token": "<|endoftext|>",
+  "additional_special_tokens": [
+    ">>TITLE<<",
+    ">>ABSTRACT<<",
+    ">>INTRODUCTION<<",
+    ">>SUMMARY<<",
+    ">>COMMENT<<",
+    ">>ANSWER<<",
+    ">>QUESTION<<",
+    ">>DOMAIN<<",
+    ">>PREFIX<<",
+    ">>SUFFIX<<",
+    ">>MIDDLE<<"
+  ],
   "eos_token": "<|endoftext|>",
-  "pad_token": "<|endoftext|>",
-  "unk_token": "<|endoftext|>"
+  "pad_token": "<|endoftext|>"
 }

tokenizer_config.json

@@ -1,9 +1,7 @@
 {
   "add_prefix_space": false,
-  "bos_token": "<|endoftext|>",
   "clean_up_tokenization_spaces": true,
   "eos_token": "<|endoftext|>",
-  "model_max_length": 1024,
-  "tokenizer_class": "GPT2Tokenizer",
-  "unk_token": "<|endoftext|>"
+  "model_max_length": 2048,
+  "tokenizer_class": "PreTrainedTokenizerFast"
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fab13d37bec87b543044a9fa78ecc3dc502fd351f1d3ea58bf216ad1ede9b749
+oid sha256:47519cda35bbccb397bbf8e0f0df3d5fc383699039b7d6e00573e7937e342dfd
 size 4027