All relevant files added

example_inputs/task1_example.txt

@@ -0,0 +1,2 @@
+þeta er texi inniheldur möuelga villur.
+Kversu vel ganga leiretta þenna texta

example_inputs/task2_example.jsonl

@@ -0,0 +1,2 @@
+{"a": "Þetta er texti sem inniheldur mögulega einhverjar villur.", "b": "þeta er texi inniheldur möuelga villur."}
+{"a": "Kversu vel ganga velja mili þessarra texta", "b": "Hversu vel ætli gangi að velja milli þessara texta?"}

example_inputs/task3_example.txt

@@ -0,0 +1,2 @@
+þeta er texi inniheldur möuelga villur.
+Kversu vel ganga leiretta þenna texta

example_outputs/task1_example.txt

@@ -0,0 +1,8 @@
+þeta er texi inniheldur möuelga villur.
+- 'þeta' á undan 'er' væri betur ritað sem 'Þetta'.
+- 'texi' beint á eftir 'er' og á undan 'inniheldur' væri betur ritað sem 'texti'.
+- 'möuelga' beint á eftir 'inniheldur' og á undan 'villur' væri betur ritað sem 'mögulega'.
+
+Kversu vel ganga leiretta þenna texta
+- 'leiretta' beint á eftir 'gegna' og á undan 'þenna' væri betur ritað sem 'leirrétta'.
+

example_outputs/task2_example.txt

@@ -0,0 +1,2 @@
+B
+B

example_outputs/task3_example.txt

@@ -0,0 +1,2 @@
+Þetta er texti sem inniheldur mögulega villur.
+Kversu vel gengur að þýða þennan texta.

gpt-sw3-model/config.json

@@ -0,0 +1,37 @@
+{
+  "_name_or_path": "/p/scratch/trustllm/ingvarssonjuto1/GPT-SW3/checkpoints_run_6/unwrapped_step_8192",
+  "activation_function": "gelu",
+  "apply_query_key_layer_scaling": true,
+  "architectures": [
+    "GPT2LMHeadModel"
+  ],
+  "attn_pdrop": 0.1,
+  "bos_token_id": 1,
+  "embd_pdrop": 0.1,
+  "eos_token_id": 1,
+  "initializer_range": 0.01,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "gpt2",
+  "n_ctx": 2048,
+  "n_embd": 4096,
+  "n_head": 32,
+  "n_inner": 16384,
+  "n_layer": 32,
+  "n_positions": 2048,
+  "normalize_attention_scores": true,
+  "pad_token_id": 0,
+  "reorder_and_upcast_attn": false,
+  "resid_pdrop": 0.1,
+  "scale_attn_by_inverse_layer_idx": false,
+  "scale_attn_weights": true,
+  "summary_activation": null,
+  "summary_first_dropout": 0.1,
+  "summary_proj_to_labels": true,
+  "summary_type": "cls_index",
+  "summary_use_proj": true,
+  "tokenizer_class": "GPTSw3Tokenizer",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.38.1",
+  "use_cache": false,
+  "vocab_size": 64000
+}

gpt-sw3-model/generation_config.json

@@ -0,0 +1,8 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 1,
+  "pad_token_id": 0,
+  "transformers_version": "4.38.1",
+  "use_cache": false
+}

gpt-sw3-model/pytorch_model-00001-of-00003.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3b5613e9bb03e311bf50ed0ecce4594159dbf1edaf228566b5a8dd3ae5c46bd2
+size 4971486034

gpt-sw3-model/pytorch_model-00002-of-00003.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a7fdf1eeefc27df8163b1cf635a8d65ee17eda4f5ca343ad6638a5d9d7b6e1b2
+size 4967435972

gpt-sw3-model/pytorch_model-00003-of-00003.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7d6e0c04f3c19fa3eb66071f588d1d15424b3938570dbe1632248d723a97100
+size 4014894443

gpt-sw3-model/pytorch_model.bin.index.json

@@ -0,0 +1,396 @@
+{
+  "metadata": {
+    "total_size": 13953679360
+  },
+  "weight_map": {
+    "lm_head.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.0.attn.c_attn.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.0.attn.c_attn.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.0.attn.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.0.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.0.ln_1.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.0.ln_1.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.0.ln_2.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.0.ln_2.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.0.mlp.c_fc.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.0.mlp.c_fc.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.0.mlp.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.0.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.1.attn.c_attn.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.1.attn.c_attn.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.1.attn.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.1.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.1.ln_1.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.1.ln_1.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.1.ln_2.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.1.ln_2.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.1.mlp.c_fc.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.1.mlp.c_fc.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.1.mlp.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.1.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.10.attn.c_attn.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.10.attn.c_attn.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.10.attn.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.10.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.10.ln_1.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.10.ln_1.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.10.ln_2.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.10.ln_2.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.10.mlp.c_fc.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.10.mlp.c_fc.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.10.mlp.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.10.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.11.attn.c_attn.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.11.attn.c_attn.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.11.attn.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.11.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.11.ln_1.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.11.ln_1.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.11.ln_2.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.11.ln_2.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.11.mlp.c_fc.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.11.mlp.c_fc.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.11.mlp.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.11.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.12.attn.c_attn.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.12.attn.c_attn.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.12.attn.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.12.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.12.ln_1.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.12.ln_1.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.12.ln_2.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.12.ln_2.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.12.mlp.c_fc.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.12.mlp.c_fc.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.12.mlp.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.12.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.13.attn.c_attn.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.13.attn.c_attn.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.13.attn.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.13.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.13.ln_1.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.13.ln_1.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.13.ln_2.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.13.ln_2.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.13.mlp.c_fc.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.13.mlp.c_fc.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.13.mlp.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.13.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.14.attn.c_attn.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.14.attn.c_attn.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.14.attn.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.14.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.14.ln_1.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.14.ln_1.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.14.ln_2.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.14.ln_2.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.14.mlp.c_fc.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.14.mlp.c_fc.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.14.mlp.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.14.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.15.attn.c_attn.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.15.attn.c_attn.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.15.attn.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.15.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.15.ln_1.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.15.ln_1.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.15.ln_2.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.15.ln_2.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.15.mlp.c_fc.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.15.mlp.c_fc.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.15.mlp.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.15.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.16.attn.c_attn.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.16.attn.c_attn.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.16.attn.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.16.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.16.ln_1.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.16.ln_1.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.16.ln_2.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.16.ln_2.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.16.mlp.c_fc.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.16.mlp.c_fc.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.16.mlp.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.16.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.17.attn.c_attn.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.17.attn.c_attn.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.17.attn.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.17.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.17.ln_1.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.17.ln_1.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.17.ln_2.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.17.ln_2.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.17.mlp.c_fc.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.17.mlp.c_fc.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.17.mlp.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.17.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.18.attn.c_attn.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.18.attn.c_attn.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.18.attn.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.18.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.18.ln_1.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.18.ln_1.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.18.ln_2.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.18.ln_2.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.18.mlp.c_fc.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.18.mlp.c_fc.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.18.mlp.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.18.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.19.attn.c_attn.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.19.attn.c_attn.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.19.attn.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.19.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.19.ln_1.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.19.ln_1.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.19.ln_2.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.19.ln_2.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.19.mlp.c_fc.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.19.mlp.c_fc.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.19.mlp.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.19.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.2.attn.c_attn.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.2.attn.c_attn.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.2.attn.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.2.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.2.ln_1.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.2.ln_1.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.2.ln_2.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.2.ln_2.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.2.mlp.c_fc.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.2.mlp.c_fc.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.2.mlp.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.2.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.20.attn.c_attn.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.20.attn.c_attn.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.20.attn.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.20.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.20.ln_1.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.20.ln_1.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.20.ln_2.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.20.ln_2.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.20.mlp.c_fc.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.20.mlp.c_fc.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.20.mlp.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.20.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.21.attn.c_attn.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.21.attn.c_attn.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.21.attn.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.21.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.21.ln_1.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.21.ln_1.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.21.ln_2.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.21.ln_2.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.21.mlp.c_fc.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.21.mlp.c_fc.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.21.mlp.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.21.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.22.attn.c_attn.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.22.attn.c_attn.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.22.attn.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.22.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.22.ln_1.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.22.ln_1.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.22.ln_2.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.22.ln_2.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.22.mlp.c_fc.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.22.mlp.c_fc.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.22.mlp.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.22.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.23.attn.c_attn.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.23.attn.c_attn.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.23.attn.c_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.23.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.23.ln_1.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.23.ln_1.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.23.ln_2.bias": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.23.ln_2.weight": "pytorch_model-00002-of-00003.bin",
+    "transformer.h.23.mlp.c_fc.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.23.mlp.c_fc.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.23.mlp.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.23.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.24.attn.c_attn.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.24.attn.c_attn.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.24.attn.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.24.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.24.ln_1.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.24.ln_1.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.24.ln_2.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.24.ln_2.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.24.mlp.c_fc.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.24.mlp.c_fc.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.24.mlp.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.24.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.25.attn.c_attn.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.25.attn.c_attn.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.25.attn.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.25.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.25.ln_1.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.25.ln_1.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.25.ln_2.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.25.ln_2.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.25.mlp.c_fc.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.25.mlp.c_fc.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.25.mlp.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.25.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.26.attn.c_attn.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.26.attn.c_attn.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.26.attn.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.26.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.26.ln_1.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.26.ln_1.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.26.ln_2.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.26.ln_2.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.26.mlp.c_fc.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.26.mlp.c_fc.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.26.mlp.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.26.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.27.attn.c_attn.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.27.attn.c_attn.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.27.attn.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.27.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.27.ln_1.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.27.ln_1.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.27.ln_2.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.27.ln_2.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.27.mlp.c_fc.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.27.mlp.c_fc.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.27.mlp.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.27.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.28.attn.c_attn.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.28.attn.c_attn.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.28.attn.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.28.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.28.ln_1.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.28.ln_1.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.28.ln_2.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.28.ln_2.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.28.mlp.c_fc.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.28.mlp.c_fc.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.28.mlp.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.28.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.29.attn.c_attn.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.29.attn.c_attn.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.29.attn.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.29.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.29.ln_1.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.29.ln_1.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.29.ln_2.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.29.ln_2.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.29.mlp.c_fc.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.29.mlp.c_fc.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.29.mlp.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.29.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.3.attn.c_attn.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.3.attn.c_attn.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.3.attn.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.3.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.3.ln_1.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.3.ln_1.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.3.ln_2.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.3.ln_2.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.3.mlp.c_fc.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.3.mlp.c_fc.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.3.mlp.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.3.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.30.attn.c_attn.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.30.attn.c_attn.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.30.attn.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.30.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.30.ln_1.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.30.ln_1.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.30.ln_2.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.30.ln_2.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.30.mlp.c_fc.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.30.mlp.c_fc.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.30.mlp.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.30.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.31.attn.c_attn.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.31.attn.c_attn.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.31.attn.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.31.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.31.ln_1.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.31.ln_1.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.31.ln_2.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.31.ln_2.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.31.mlp.c_fc.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.31.mlp.c_fc.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.31.mlp.c_proj.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.31.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.h.4.attn.c_attn.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.4.attn.c_attn.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.4.attn.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.4.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.4.ln_1.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.4.ln_1.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.4.ln_2.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.4.ln_2.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.4.mlp.c_fc.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.4.mlp.c_fc.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.4.mlp.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.4.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.5.attn.c_attn.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.5.attn.c_attn.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.5.attn.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.5.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.5.ln_1.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.5.ln_1.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.5.ln_2.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.5.ln_2.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.5.mlp.c_fc.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.5.mlp.c_fc.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.5.mlp.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.5.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.6.attn.c_attn.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.6.attn.c_attn.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.6.attn.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.6.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.6.ln_1.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.6.ln_1.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.6.ln_2.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.6.ln_2.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.6.mlp.c_fc.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.6.mlp.c_fc.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.6.mlp.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.6.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.7.attn.c_attn.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.7.attn.c_attn.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.7.attn.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.7.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.7.ln_1.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.7.ln_1.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.7.ln_2.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.7.ln_2.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.7.mlp.c_fc.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.7.mlp.c_fc.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.7.mlp.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.7.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.8.attn.c_attn.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.8.attn.c_attn.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.8.attn.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.8.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.8.ln_1.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.8.ln_1.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.8.ln_2.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.8.ln_2.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.8.mlp.c_fc.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.8.mlp.c_fc.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.8.mlp.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.8.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.9.attn.c_attn.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.9.attn.c_attn.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.9.attn.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.9.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.9.ln_1.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.9.ln_1.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.9.ln_2.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.9.ln_2.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.9.mlp.c_fc.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.9.mlp.c_fc.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.9.mlp.c_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "transformer.h.9.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.ln_f.bias": "pytorch_model-00003-of-00003.bin",
+    "transformer.ln_f.weight": "pytorch_model-00003-of-00003.bin",
+    "transformer.wpe.weight": "pytorch_model-00001-of-00003.bin",
+    "transformer.wte.weight": "pytorch_model-00001-of-00003.bin"
+  }
+}

requirements.txt

@@ -0,0 +1,3 @@
+torch==2.2.1
+transformers==4.38.2
+tqdm==4.66.2

run_model.py

@@ -0,0 +1,180 @@
+"""This script runs the trained model on data and saves the predictions to a file."""
+
+from transformers import AutoModelForCausalLM, AutoTokenizer
+import torch
+import logging
+import random
+import tqdm
+import json
+import argparse
+
+# Set the logging level to info
+logging.basicConfig(level=logging.INFO)
+
+# Set the device to GPU if available
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+logging.info(f"Device: {device}")
+
+# Prompts for the different tasks
+START_PROMPT_TASK1 = "Hér er texti sem ég vil að þú skoðir vel og vandlega. Þú skalt skoða hvert einasta orð, orðasamband, og setningu og meta hvort þér finnist eitthvað athugavert, til dæmis hvað varðar málfræði, stafsetningu, skringilega merkingu og svo framvegis.\nHér er textinn:\n\n"
+END_PROMPT_TASK1 = "Sérðu eitthvað sem mætti betur fara í textanum? Búðu til lista af öllum slíkum tilvikum þar sem hver lína tilgreinir hver villan er, hvar hún er, og hvað væri gert í staðinn fyrir villuna.\n\n"
+
+START_PROMPT_TASK2 = "Hér er texti sem ég vil að þú skoðir vel og vandlega. Þú skalt skoða hvert einasta orð, orðasamband, og setningu og meta hvort þér finnist eitthvað athugavert, til dæmis hvað varðar málfræði, stafsetningu, skringilega merkingu og svo framvegis.Ég er með tvær útgáfur af textanum, A og B, og önnur þeirra gæti verið betri en hin á einhvern hátt, t.d. hvað varðar stafsetningu, málfræði o.s.frv.\nHér er texti A:\n\n"
+MIDDLE_PROMPT_TASK2 = "Hér er texti B:\n\n"
+END_PROMPT_TASK2 = "Hvorn textann líst þér betur á?\n\n"
+
+START_PROMPT_TASK3 = "Hér er texti sem ég vil að þú skoðir vel og vandlega. Þú skalt skoða hvert einasta orð, orðasamband, og setningu og meta hvort þér finnist eitthvað athugavert, til dæmis hvað varðar málfræði, stafsetningu, skringilega merkingu og svo framvegis.\nHér er textinn:\n\n"
+END_PROMPT_TASK3 = "Reyndu nú að laga textann þannig að hann líti betur út, eins og þér finnst best við hæfi.\n\n"
+
+START_PROMPT_TASK = {
+    1: START_PROMPT_TASK1,
+    2: START_PROMPT_TASK2,
+    3: START_PROMPT_TASK3,
+}
+END_PROMPT_TASK = {1: END_PROMPT_TASK1, 2: END_PROMPT_TASK2, 3: END_PROMPT_TASK3}
+
+SEP = "\n\n"
+
+
+def set_seed(seed):
+    """Set the random seed for reproducibility."""
+    torch.manual_seed(seed)
+    if torch.cuda.is_available():
+        torch.cuda.manual_seed_all(seed)
+    torch.backends.cudnn.deterministic = True
+    torch.backends.cudnn.benchmark = False
+    random.seed(seed)
+
+
+def tokenize_data(tokenizer, data, task, max_length):
+    """Tokenize the data and return the input_ids and attention_mask."""
+    tokenized_start = tokenizer(START_PROMPT_TASK[task])["input_ids"]
+    tokenized_end = tokenizer(END_PROMPT_TASK[task])["input_ids"]
+    if task == 2:
+        tokenized_middle = tokenizer(MIDDLE_PROMPT_TASK2)["input_ids"]
+
+    # Tokenize the data
+    tokenized_data = []
+    if task == 1 or task == 3:
+        for sentence in data:
+            tokenized_sentence = tokenizer(sentence + SEP)["input_ids"]
+
+            # Concatenate the tokenized data
+            concatted_data = (
+                [tokenizer.bos_token_id]
+                + tokenized_start
+                + tokenized_sentence
+                + tokenized_end
+            )
+
+            # Truncate the data
+            concatted_data = concatted_data[:max_length]
+
+            tokenized_data.append(concatted_data)
+    elif task == 2:
+        for line in data:
+            data_a = line["a"]
+            data_b = line["b"]
+            tokenized_sentence_a = tokenizer(data_a + SEP)["input_ids"]
+            tokenized_sentence_b = tokenizer(data_b + SEP)["input_ids"]
+
+            # Concatenate the tokenized data
+            concatted_data = (
+                [tokenizer.bos_token_id]
+                + tokenized_start
+                + tokenized_sentence_a
+                + tokenized_middle
+                + tokenized_sentence_b
+                + tokenized_end
+            )
+
+            # Truncate the data
+            concatted_data = concatted_data[:max_length]
+
+            tokenized_data.append(concatted_data)
+
+    return tokenized_data
+
+
+def run_model_on_data(model_path, tokenizer_name, arguments):
+    """Run the model on the data and save the predictions to a file."""
+    # Load the model and tokenizer
+    model = AutoModelForCausalLM.from_pretrained(model_path, torch_dtype=torch.bfloat16)
+    model.to(device)
+    model.eval()
+    tokenizer = AutoTokenizer.from_pretrained(tokenizer_name)
+
+    # Load the data
+    if arguments.task == 1 or arguments.task == 3:
+        with open(arguments.input_file, "r") as file:
+            data = file.read().splitlines()
+    elif arguments.task == 2:
+        with open(arguments.input_file, "r") as file:
+            data = file.read().splitlines()
+            data = [json.loads(line) for line in data]
+
+    # Tokenize the data
+    data_tokenized = tokenize_data(
+        tokenizer, data, arguments.task, tokenizer.model_max_length
+    )
+    logging.info(f"Number of examples: {len(data_tokenized)}")
+
+    # Run the model on the data
+    predictions = []
+    progress_bar = tqdm.tqdm(total=len(data_tokenized), desc="Running model on data")
+
+    for input_ids in data_tokenized:
+        progress_bar.update(1)
+
+        # Generate the predictions
+        with torch.cuda.amp.autocast(dtype=torch.bfloat16):
+            input_ids_tensor = torch.tensor(input_ids).unsqueeze(0).to(device)
+            output = model.generate(
+                input_ids=input_ids_tensor, max_new_tokens=500, num_return_sequences=1
+            )
+
+            # Only get the part of the prediction that was generated
+            prediction = tokenizer.decode(
+                output[0][len(input_ids) :], skip_special_tokens=True
+            )
+            predictions.append(prediction)
+
+    progress_bar.close()
+
+    # Save the predictions to a file
+    with open(arguments.output_file, "w") as file:
+        if arguments.task == 1:
+            # We want to include the original text in the output file
+            counter = 0
+            for prediction in predictions:
+                file.write(data[counter] + "\n")
+                file.write(prediction.split("\n\n")[0] + "\n\n")
+                counter += 1
+        else:
+            for prediction in predictions:
+                file.write(prediction.split("\n\n")[0] + "\n")
+
+    logging.info(f"Predictions written to file: {arguments.output_file}")
+
+
+if __name__ == "__main__":
+    # Parse the arguments
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--task", type=int, help="The task type (1, 2, or 3)")
+    parser.add_argument(
+        "--input-file",
+        type=str,
+        help="The path to the input file with data to be corrected",
+    )
+    parser.add_argument(
+        "--output-file",
+        type=str,
+        help="The path to the output file where the corrected data will be saved",
+    )
+    args = parser.parse_args()
+
+    model_path = "./gpt-sw3-model"
+    tokenizer_name = "AI-Sweden-Models/gpt-sw3-6.7b"
+
+    set_seed(42)
+    run_model_on_data(model_path, tokenizer_name, args)