refactor: only models here, training pipeline archive on gh

README.md

@@ -15,35 +15,249 @@ library_name: stanza
 
 # Retraining Stanza to optimize dependency parsing on a diachronic Swedish corpus
 
-This repository contains forked code from the official Stanza github, with scripts that help prepare for and train models on different combinations of treebanks relevant to historical Swedish. 
+This repository contains Stanza BiLSTM models retrained on different combinations of UD treebanks relevant to historical Swedish. The models prefixed `conll17_` are trained with static embeddings, and the models prefixed `transformer_` are trained with dynamic embeddings from the transformer encoder "KBLab/bert-base-swedish-cased".
 
-## Guide
+## LAS Scores
 
-Dev/test for all models is 10/90 of our human validated gold sentences (https://github.com/alanev52/Diachronic_Treebanks_DigPhil/tree/main/parsed_data/validated)
+LAS scores for the models are computed against a gold set of 109 manually annoted sentences divided into five different periods.
+For the models trained on static vectors, only the overall test scores is given:
 
-Example work flow, training a model with pretrained vectors from kubhist2 1880 incremental and training data from Swedish and Bokmål treebanks, as well as our own machine-parsed diachronic corpus:
+### Models with static vector embeddings (`conll17.pt`)
 
-```
-python prepare-train-val-test.py sv diachron bm
+| Languages                      | LAS   |
+|--------------------------------|-------|
+| **Swedish (with diachronic)**  | **61.95** |
+| Icelandic (PUD)                | 61.49 |
+| German (LIT)                   | 61.43 |
+| Icelandic (GC)                 | 61.43 |
+| Bokmaal, Danish                | 60.13 |
+| Nynorsk                        | 50.46 |
+| *Swedish (without diachronic)* | *50.34* |
+| Icelandic (Modern)             | 46.47 |
+| Bokmaal                        | 45.96 |
+| Icelandic (IcePaHC)            | 44.60 |
+
+For the transformer-fed models, more fine-grained scores on each period are given as a histogram. The model `transformer_seen_gold_no_silver.pt` was given the gold set during training and hence has no score, but is intuitively the best model. As a benchmark, an "out-of-the-box" Stanza trained only on Talbanken is given.
 
-source scripts/config_alvis.sh
+<img src="plot/las_f1_scores_comparison.png" width="40%">
 
-python -m stanza.utils.datasets.prepare_depparse_treebank UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
+## Inference
 
-python -m stanza.utils.training.run_depparse UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt --batch_size 32 --dropout 0.33
+Example for how the models can be run:
 ```
+import os
+from pathlib import Path
+import stanza
+from stanza.utils.conll import CoNLL
+import time
+import gc
+import torch
+from tqdm import tqdm
 
-All of the above can be done with a single command by using:
+if torch.xpu.is_available():
+    device = torch.device("xpu")
+elif torch.cuda.is_available():
+    device = torch.device("cuda")
+else:
+    device = torch.device("cpu")
 
-```
-./make_new_model.sh {vectors} {language codes}
-```
+###################################################################################################
+##### SETTINGS ####################################################################################
+###################################################################################################
 
-which for the example becomes:
+vanilla = False
 
+conllu_in_dir = Path("YOUR/CONLLUS/")
+conllu_out_dir = Path("OUT/DIR")
+
+depparse_model_path = "transformer_seen_gold_no_silver.pt"
+
+os.makedirs(conllu_out_dir, exist_ok=True)
+
+##### PREPARE #####################################################################################
+
+if vanilla:
+    print("Using vanilla Swedish Stanza pipeline with default models.")
+    nlp = stanza.Pipeline(
+        "sv",
+        processors="tokenize,pos,lemma,depparse",
+        tokenize_pretokenized=True, # to keep original tokens 
+        use_gpu=True,
+        pos_batch_size=3000,
+        package=None,
+        device=device,
+    )
+
+else:
+    nlp = stanza.Pipeline(
+        "sv",
+        processors="tokenize,pos,lemma,depparse",
+        tokenize_pretokenized=True, # to keep original tokens 
+        use_gpu=True,
+        pos_batch_size=3000,
+        package=None,
+        download_method=None,
+        depparse_model_path=depparse_model_path, # depparse_model_path becomes model_path inside the depparse processor config. explicit model_path is to control where normal Stanza models are stored
+        device=device,
+    )
+
+##### INFERENCE ###################################################################################
+
+t0 = time.time()
+total_sentences = 0
+
+for fname in tqdm(sorted(os.listdir(conllu_in_dir)), desc="Files"):
+    if not fname.endswith(".conllu"):
+        continue
+
+    in_path = conllu_in_dir / fname
+    out_path = conllu_out_dir / fname
+    
+    if out_path.exists():
+        print(f"File already parsed, skipping: {out_path}")
+        continue
+
+    doc = CoNLL.conll2doc(in_path)
+
+    total_sentences += len(doc.sentences)
+
+    with torch.inference_mode():
+        parsed_doc = nlp(doc)
+
+    with open(out_path, "w", encoding="utf-8") as f:
+        CoNLL.write_doc2conll(parsed_doc, f)
+
+    # Release per-file objects and ask both Python and the torch allocator to
+    # reclaim memory that is no longer needed.
+    del doc
+    del parsed_doc
+    gc.collect()
+    if device.type == "cuda":
+        torch.cuda.empty_cache()
+        memory_max = torch.cuda.max_memory_allocated()
+        print(f"Max memory allocated so far: {memory_max / (1024 ** 3):.2f} GB")
+    elif device.type == "xpu":
+        torch.xpu.empty_cache()
+        memory_max = torch.xpu.max_memory_allocated()
+        print(f"Max memory allocated so far: {memory_max / (1024 ** 3):.2f} GB")
+
+t = time.time() - t0
+
+###################################################################################################
+
+if device.type == "cuda":
+    memory_max = torch.cuda.max_memory_allocated()
+    torch.cuda.reset_peak_memory_stats()
+elif device.type == "xpu":
+    memory_max = torch.xpu.max_memory_allocated()
+    torch.xpu.reset_peak_memory_stats()
+else:
+    memory_max = 0
+
+print(
+    "\nFinished parsing.\n"
+    f"Total sentences parsed: {total_sentences}\n"
+    f"Total time: {t / 60:.2f} minutes\n"
+    f"Max memory allocated: {memory_max / (1024 ** 3):.2f} GB\n"
+)
 ```
-./make_new_model.sh diachronic.pt sv diachron bm
-```
 
-## Results
+## Training args
+
+Full list of training args:
 
+```
+batch_size: 32
+bert_finetune: False
+bert_finetune_layers: None
+bert_hidden_layers: 4
+bert_learning_rate: 1.0
+bert_model: KBLab/bert-base-swedish-cased
+bert_start_finetuning: 200
+bert_warmup_steps: 200
+bert_weight_decay: 0.0
+beta2: 0.999
+char: True
+char_emb_dim: 100
+char_hidden_dim: 400
+char_num_layers: 1
+char_rec_dropout: 0
+charlm: True
+charlm_backward_file: /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt
+charlm_forward_file: /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt
+charlm_save_dir: saved_models/charlm
+charlm_shorthand: sv_conll17
+checkpoint: True
+checkpoint_interval: 500
+checkpoint_save_name: None
+continue_from: None
+data_dir: data/depparse
+deep_biaff_hidden_dim: 400
+deep_biaff_output_dim: 160
+device: xpu:0
+distance: True
+dropout: 0.33
+eval_file: /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
+eval_interval: 100
+gold_labels: True
+hidden_dim: 400
+lang: sv
+linearization: True
+log_norms: False
+log_step: 20
+lora_alpha: 128
+lora_dropout: 0.1
+lora_modules_to_save: []
+lora_rank: 64
+lora_target_modules: ['query', 'value', 'output.dense', 'intermediate.dense']
+lr: 2.0
+max_grad_norm: 1.0
+max_steps: 50000
+max_steps_before_stop: 2000
+mode: train
+model_type: graph
+num_layers: 3
+optim: adadelta
+output_file: None
+output_latex: False
+pretrain: True
+pretrain_max_vocab: 250000
+rec_dropout: 0
+reversed: False
+sample_train: 1.0
+save_dir: saved_models/depparse
+save_name: {shorthand}_{embedding}_parser.pt
+second_batch_size: None
+second_bert_learning_rate: 0.001
+second_lr: 0.0002
+second_optim: adam
+second_optim_start_step: 10000
+second_warmup_steps: 200
+seed: 1234
+shorthand: sv_diachronic
+silver_file: None
+silver_weight: 0.5
+tag_emb_dim: 50
+train_file: /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
+train_size: None
+transformed_dim: 125
+transition_embedding_dim: 20
+transition_hidden_dim: 20
+transition_merge_hidden_dim: 200
+transition_subtree_combination: SubtreeCombination.NONE
+transition_subtree_nonlinearity: none
+use_arc_embedding: False
+use_peft: False
+use_ufeats: True
+use_upos: True
+use_xpos: True
+wandb: False
+wandb_name: None
+weight_decay: 1e-05
+word_cutoff: 7
+word_dropout: 0.33
+word_emb_dim: 75
+wordvec_dir: /home/urdatorn/stanza_resources/sv/pretrain
+wordvec_file: None
+wordvec_pretrain_file: /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt
+```

parsed/old-conll17-1/svediakorp-letter141673-Stalhammar.conllu → conll17_baseline_sv_only.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0e4779224fe6e1d1580faa201af29e0500110596035652fc788e4a5b0131dd9b
-size 4846318
+oid sha256:50aee8f0eab8b4b0e28958c0e428f1aad29cf118e4320332862b8701fc6dd6e3
+size 141291582

parsed/old-conll17-1/svediakorp-letter141673-Stalhammar_preprocessed.conllu → conll17_bm.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5dbe992db0d973c529d09bfd34bf1c647db3c17aed5e51039e044e859323b472
-size 16025
+oid sha256:4b0848c6832ca155fcfbf78040530aeb27598562cda430821afddbead91ea0b1
+size 148183635

eval/gold_as_val.conllu → conll17_bm_dk.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f4e442db9582dfc10263eb2cf6cbed4719194ea0f6a1ab0ff82fa652b7a414c7
-size 111668
+oid sha256:ab11092dd5795e51add850b4aa325a474b72f374b93df53a78fd247ad456497c
+size 150762047

parsed/old-conll17-1/svediakorp-sec1018-spf175.conllu → conll17_de_lit.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2d50cf732cd496a2e203b2fd7e799147d2b6e6919711b84d0d4b702523833dc0
-size 753506
+oid sha256:b84862d0f728ba53e2076208a94d5e90bd1dc7e0f2884a12db3f21ffcaf073dc
+size 146969172

conll17_icepahc.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f59e82a5e94514bdf086d19931a26f5cd6274c6ad29c667f3366e93be0bec585
+size 151645170

conll17_is-modern.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e02dffc9ae0d419373d1ba597f4d7e0459d70d6909e18211d40c33472e5a6d46
+size 148251198

conll17_isPUD-pahc-gc.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5fe5f0e97a6a4e257b124f8b229c59cc61d58124512421d56795ac6ffb67359e
+size 146969174

conll17_isPUD.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dcdc23474d269b8f4dba7d246f184b42a446cae7294f1134121cf5da8511efa6
+size 146117067

conll17_nn.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d74ed335831d36ee6d4569d1496a62860779c378ed391a693a8aa09159f388e
+size 147986243

conll17_sv_diachron.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a67914c27dcc9cdc33034cdbebed9eec4c55b10ab2aeffc5f5d6053a96c5107a
+size 145283635

digphil_inference_pipeline.py

@@ -1,114 +0,0 @@
-import os
-from pathlib import Path
-import stanza
-from stanza.utils.conll import CoNLL
-import time
-import gc
-import torch
-from tqdm import tqdm
-
-if torch.xpu.is_available():
-  device = torch.device("xpu")
-elif torch.cuda.is_available():
-  device = torch.device("cuda")
-
-###################################################################################################
-##### SETTINGS ####################################################################################
-###################################################################################################
-
-vanilla = False
-
-conllu_in_dir = Path("/home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated")
-#conllu_out_dir = Path("parsed/testset_talbanken_not_seen_gold")
-conllu_out_dir = Path("parsed/testset_transformer_not_seen_gold_no_silver")
-
-depparse_model_path = "saved_models/depparse/transformer_not_seen_gold_no_silver.pt"
-
-os.makedirs(conllu_out_dir, exist_ok=True)
-
-##### PREPARE #####################################################################################
-
-if vanilla:
-    print("Using vanilla Swedish Stanza pipeline with default models.")
-    nlp = stanza.Pipeline(
-        "sv",
-        processors="tokenize,pos,lemma,depparse",
-        tokenize_pretokenized=True, # to keep original tokens 
-        use_gpu=True,
-        pos_batch_size=3000,
-        package=None,
-        device=device,
-    )
-
-else:
-    nlp = stanza.Pipeline(
-        "sv",
-        processors="tokenize,pos,lemma,depparse",
-        tokenize_pretokenized=True, # to keep original tokens 
-        use_gpu=True,
-        pos_batch_size=3000,
-        package=None,
-        download_method=None,
-        depparse_model_path=depparse_model_path, # depparse_model_path becomes model_path inside the depparse processor config. explicit model_path is to control where normal Stanza models are stored
-        device=device,
-    )
-
-##### INFERENCE ###################################################################################
-
-t0 = time.time()
-total_sentences = 0
-
-for fname in tqdm(sorted(os.listdir(conllu_in_dir)), desc="Files"):
-    if not fname.endswith(".conllu"):
-        continue
-
-    in_path = conllu_in_dir / fname
-    out_path = conllu_out_dir / fname
-    
-    if out_path.exists():
-        print(f"File already parsed, skipping: {out_path}")
-        continue
-
-    doc = CoNLL.conll2doc(in_path)
-
-    total_sentences += len(doc.sentences)
-
-    with torch.inference_mode():
-        parsed_doc = nlp(doc)
-
-    with open(out_path, "w", encoding="utf-8") as f:
-        CoNLL.write_doc2conll(parsed_doc, f)
-
-    # Release per-file objects and ask both Python and the torch allocator to
-    # reclaim memory that is no longer needed.
-    del doc
-    del parsed_doc
-    gc.collect()
-    if device.type == "cuda":
-        torch.cuda.empty_cache()
-        memory_max = torch.cuda.max_memory_allocated()
-        print(f"Max memory allocated so far: {memory_max / (1024 ** 3):.2f} GB")
-    elif device.type == "xpu":
-        torch.xpu.empty_cache()
-        memory_max = torch.xpu.max_memory_allocated()
-        print(f"Max memory allocated so far: {memory_max / (1024 ** 3):.2f} GB")
-
-t = time.time() - t0
-
-###################################################################################################
-
-if device.type == "cuda":
-    memory_max = torch.cuda.max_memory_allocated()
-    torch.cuda.reset_peak_memory_stats()
-elif device.type == "xpu":
-    memory_max = torch.xpu.max_memory_allocated()
-    torch.xpu.reset_peak_memory_stats()
-else:
-    memory_max = 0
-
-print(
-    "\nFinished parsing.\n"
-    f"Total sentences parsed: {total_sentences}\n"
-    f"Total time: {t / 60:.2f} minutes\n"
-    f"Max memory allocated: {memory_max / (1024 ** 3):.2f} GB\n"
-)

eval/execute_eval.py

@@ -1,273 +0,0 @@
-from mod_conll18_ud_eval import load_conllu, evaluate
-from collections import defaultdict
-from pathlib import Path
-import pyconll
-import sys
-import os
-import numpy as np
-
-ROOT = Path.cwd()
-print(ROOT)
-gold_dir = Path("/home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated") # mapp med vårt manuellt annoterade guldstandardset
-val_file = ROOT / "eval/gold_as_val.conllu"  # sentences used during validation - must be excluded
-
-def load_val_sentence_ids(val_file_path):
-    """Load sentence IDs from the validation file to exclude from evaluation."""
-    val_ids = set()
-    if not os.path.exists(val_file_path):
-        print(f"Warning: validation file {val_file_path} not found. No sentences will be excluded.")
-        return val_ids
-    val_data = pyconll.load_from_file(str(val_file_path))
-    for sentence in val_data:
-        val_ids.add(sentence.id)
-    print(f"Loaded {len(val_ids)} sentence IDs to exclude from evaluation (used in validation).")
-    return val_ids
-
-def filter_gold_file(gold_file_path, excluded_ids):
-    """Create a filtered gold file excluding sentences used during validation.
-    Returns path to filtered file and the number of excluded sentences."""
-    gold_data = pyconll.load_from_file(gold_file_path)
-    filtered_sentences = [s for s in gold_data if s.id not in excluded_ids]
-    n_excluded = len(gold_data) - len(filtered_sentences)
-
-    if n_excluded > 0:
-        print(f"  Excluded {n_excluded} validation sentences from {os.path.basename(gold_file_path)}")
-
-    # Write filtered data to a temp file
-    filtered_path = gold_file_path.replace(".conllu", "_eval_filtered.conllu")
-    with open(filtered_path, "w") as f:
-        for sentence in filtered_sentences:
-            f.write(sentence.conll())
-            f.write("\n\n")
-    return filtered_path, len(filtered_sentences)
-
-# Load validation sentence IDs at startup
-val_sentence_ids = load_val_sentence_ids(val_file)
-
-# Model configurations: (predicted_dir, output_filename)
-MODEL_CONFIGS = [
-    (ROOT / "parsed/testset_talbanken_not_seen_gold", "eval/scores/scores_talbanken.txt"),
-    (ROOT / "parsed/testset_transformer_silver_not_seen_gold", "eval/scores/scores_transformer_silver.txt"),
-    (ROOT / "parsed/testset_transformer_not_seen_gold_no_silver", "eval/scores/scores_transformer_no_silver.txt"),
-]
-
-def check_valid_gold(gold_file_path):
-    """ Check that the gold trees are valid """
-    gold_data = pyconll.load_from_file(gold_file_path)
-    errors = []
-    for sentence in gold_data:
-        for token in sentence:
-            if token.head is None:
-                e = ValueError(f"Invalid gold tree in file {gold_file_path}, sentence ID {sentence.id}: token {token.id} '{token.form}' has no head.")
-                errors.append(e)
-            elif token.head == token.id:
-                e = ValueError(f"Invalid gold tree in file {gold_file_path}, sentence ID {sentence.id}: token {token.id} '{token.form}' has itself as head.")
-                errors.append(e)
-        roots = [token for token in sentence if token.head == '0']
-        if len(roots) != 1:
-            e = ValueError(f"Invalid gold tree in file {gold_file_path}, sentence ID {sentence.id}: expected 1 root, found {len(roots)}.")
-            errors.append(e)
-    return errors if len(errors) > 0 else None
-
-def preprocess_system_file(system_file_path, gold_file_path):
-    """ Make a new base file with only the trees present in the gold file """
-    gold_conllu = pyconll.load_from_file(gold_file_path)
-    system_conllu = pyconll.load_from_file(system_file_path)
-    gold_ids = {sentence.id for sentence in gold_conllu}
-    filtered_sentences = [sentence for sentence in system_conllu if sentence.id in gold_ids]
-    preprocessed_file_path = system_file_path.replace(".conllu", "_preprocessed.conllu")
-    with open(preprocessed_file_path, "w") as outfile:
-        for sentence in filtered_sentences:
-            outfile.write(sentence.conll())
-            outfile.write("\n\n")
-    return preprocessed_file_path
-
-def execute_evaluation(gold_file_path, predicted_file_path):
-    # Read goldfile
-    try:
-        gold_data = load_conllu(open(gold_file_path, "r", encoding="utf-8"))
-    except Exception as e:
-        print(f"Skipping text because of error reading gold file {gold_file_path}: {e}")
-        #raise e
-        return None
-    # Read predicted file and ignore some format errors
-    try:
-        predicted_data = load_conllu(open(predicted_file_path, "r", encoding="utf-8"), ignore_invalid_format=True)
-    except Exception as e:
-        print(f"Skipping text because of error reading predicted file {predicted_file_path}: {e}")
-        #raise e
-        return None
-    # Evaluate the predictions against the gold standard
-    try:
-        results = evaluate(gold_data, predicted_data)
-    except Exception as e:
-        print(f"Error during evaluation between {gold_file_path} and {predicted_file_path}: {e}")
-        raise e
-    return results
-
-time_periods = ["1700-1750", "1750-1800", "1800-1850", "1850-1900", "1900-1950"]
-metrics = ["UPOS", "UAS", "LAS"] # the relevant metrics
-
-# To look up time period index for a sample
-sample_period = {"svediakorp-letter141673-Stalhammar": 0, "svediakorp-sec25-Runius": 0, "svediakorp-sec330-GyllenborgC_SwenskaSpratthoken": 0,
-           "svediakorp-sec277-EnbomPU_MedborgeligtSkalde": 1, "svediakorp-sec268-DulciU_VitterhetsNojen3": 1,
-           "svediakorp-sec991-spf148": 2, "svediakorp-sec252-BremerF_Teckningar1": 2, "svediakorp-sec324-GranbergPA_Enslighetsalskaren": 2,
-           "svediakorp-sec254-CederborghF_BerattelseOmJohnHall": 2, "svediakorp-sec987-spf144": 2, "svediakorp-sec988-spf145": 2,
-           "svediakorp-sec452-NyblomH_FantasierFyra": 3, "svediakorp-sec486-SchwartzMS_BellmansSkor": 3, "svediakorp-sec1102-spf259": 3,
-           "svediakorp-sec208-Anonym_DetGrasligaMordet": 3, "svediakorp-sec1063-spf220": 3,
-           "svediakorp-sec631-HasselskogN_HallaHallaGronkoping": 4, "svediakorp-sec1033-spf190": 4, "svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol": 4,
-           "svediakorp-sec613-EngstromA_StrindbergOchJag": 4, "svediakorp-sec639-HeidenstamV_Proletarfilosofiens": 4}
-
-
-def run_evaluation(predicted_dir, output_file):
-    """Run evaluation for a single model configuration."""
-    print(f"\n{'='*70}")
-    print(f"Evaluating: {predicted_dir}")
-    print(f"Output: {output_file}")
-    print(f"{'='*70}\n")
-
-    # Check if predicted directory exists
-    if not os.path.exists(predicted_dir):
-        print(f"ERROR: Predicted directory does not exist: {predicted_dir}")
-        return
-
-    # Reset per-model state
-    score_arrays = [np.zeros((4, len(metrics))), np.zeros((4, len(metrics))), np.zeros((4, len(metrics))), np.zeros((4, len(metrics))), np.zeros((4, len(metrics)))]
-    skipped_samples = []
-    sample_nsents = defaultdict()
-    processed_files = []  # Track (gold_path, pred_path) pairs for overall evaluation
-
-    with open(output_file, "w") as outfile:
-        # Loop through each of the validated samples
-        for sample in os.listdir(gold_dir):
-            sample_name = sample.split("/")[-1].replace(".conllu", "")
-
-            print("DEBUG sample_name:", sample_name)
-
-            period = sample_period.get(sample_name) # int 0-4
-            print("DEBUG period value:", period)
-
-            # Filter gold file to exclude sentences used during validation
-            gold_file_path_orig = os.path.join(gold_dir, sample)
-            gold_file_path, n_filtered_sents = filter_gold_file(gold_file_path_orig, val_sentence_ids)
-
-            # Skip if all sentences were excluded
-            if n_filtered_sents == 0:
-                print(f"Skipping {sample_name}: all sentences were used in validation")
-                skipped_samples.append(sample_name)
-                continue
-
-            # Check that gold trees are valid (using filtered file)
-            e = check_valid_gold(gold_file_path)
-            if e: # skip invalid gold files
-                for error in e:
-                    print(error)
-                skipped_samples.append(sample_name)
-                continue
-            # Always regenerate preprocessed predicted file to match filtered gold
-            unprocessed_predicted_file_path = os.path.join(predicted_dir, sample)
-            if os.path.exists(unprocessed_predicted_file_path):
-                # Use filtered gold file to preprocess predictions
-                predicted_file_path = preprocess_system_file(unprocessed_predicted_file_path, gold_file_path)
-            else:
-                print(f"Skipping {sample_name}: missing predicted file {unprocessed_predicted_file_path}")
-                skipped_samples.append(sample_name)
-                continue
-            # Score base tree (parser output) against validated tree (gold standard)
-            results = execute_evaluation(gold_file_path, predicted_file_path)
-
-            if not results: # skip invalid files
-                skipped_samples.append(sample_name)
-            # Write results to scores file and save to time period arrays
-            else:
-                # Save number of sentences in lookup dict (using filtered count)
-                sample_nsents[sample_name] = n_filtered_sents
-                # Track file paths for overall evaluation
-                processed_files.append((gold_file_path, predicted_file_path))
-                print("DEBUG period value:", period)
-                print("DEBUG time_periods:", time_periods)
-
-                outfile.write(f"Sample: {sample_name}, time period: {time_periods[period]}\n")
-                outfile.write("Metric\tPrecision\tRecall\tF1\tAligned Accuracy\n")
-                metric_n = 0 # counter for loop below
-                # Loop over each metric: UPOS, UAS, LAS etc.
-                for metric, score in results.items(): # get the metric and its 4 scores
-                    if metric in metrics: # only save the relevant metrics
-                        # Save scores weighted by the number of sentences to arrays
-                        score_arrays[period][0][metric_n] += (score.precision * sample_nsents[sample_name])
-                        score_arrays[period][1][metric_n] += (score.recall * sample_nsents[sample_name])
-                        score_arrays[period][2][metric_n] += (score.f1 * sample_nsents[sample_name])
-                        # Write the metric and its file-average scores
-                        outfile.write(f"{metric}\t{score.precision:.4f}\t{score.recall:.4f}\t{score.f1:.4f}\t")
-                        if score.aligned_accuracy is None:
-                            outfile.write(f"{score.aligned_accuracy}\n")
-                        else:
-                            outfile.write(f"{score.aligned_accuracy:.4f}\n")
-                            # Save aligned accuracy to score arrays
-                            score_arrays[period][3][metric_n] += (score.aligned_accuracy * sample_nsents[sample_name])
-                        metric_n += 1 # move to next metric in array
-                outfile.write("\n")
-        outfile.write(f"{'-'*70}\n")
-
-        # After all sample scores are written, compute average scores for each time period
-        for period, period_scores in enumerate(score_arrays):
-            # Get number of samples and sentences for the period
-            samples_per_period = [sample_name for sample_name in sample_period.keys() if sample_period[sample_name] == period and sample_name not in skipped_samples]
-            n_sents_per_period = sum([sample_nsents[sample_name] for sample_name in samples_per_period])
-            # Write
-            outfile.write(f"Average scores for time period {time_periods[period]} ({len(samples_per_period)} samples with a total of {n_sents_per_period} sentences):\n")
-            outfile.write("Metric\tPrecision\tRecall\tF1\tAligned Accuracy\n")
-            for metric_n, metric in enumerate(metrics):
-                # Normalize weighted averages by the total number of sentences for the time period
-                precision_avg = period_scores[0][metric_n] / n_sents_per_period
-                recall_avg = period_scores[1][metric_n] / n_sents_per_period
-                f1_avg = period_scores[2][metric_n] / n_sents_per_period
-                aligned_acc_avg = period_scores[3][metric_n] / n_sents_per_period
-                outfile.write(f"{metric}\t{precision_avg:.4f}\t{recall_avg:.4f}\t{f1_avg:.4f}\t{aligned_acc_avg:.4f}\n")
-            outfile.write("\n")
-
-        # Compute overall scores by concatenating all files and running single evaluation
-        valid_samples = [sample_name for sample_name in sample_period.keys() if sample_name in sample_nsents and sample_name not in skipped_samples]
-        total_sents = sum(sample_nsents[sample_name] for sample_name in valid_samples)
-
-        if len(processed_files) == 0:
-            outfile.write(f"Overall scores across all time periods ({len(valid_samples)} samples with a total of {total_sents} sentences):\n")
-            outfile.write("Metric\tPrecision\tRecall\tF1\tAligned Accuracy\n")
-            outfile.write("No valid samples available for overall evaluation.\n")
-        else:
-            # Concatenate all gold and predicted files
-            all_gold_path = os.path.join(predicted_dir, "_all_gold_combined.conllu")
-            all_pred_path = os.path.join(predicted_dir, "_all_pred_combined.conllu")
-
-            with open(all_gold_path, "w") as gold_out, open(all_pred_path, "w") as pred_out:
-                for gold_path, pred_path in processed_files:
-                    with open(gold_path) as f:
-                        gold_out.write(f.read())
-                    with open(pred_path) as f:
-                        pred_out.write(f.read())
-
-            # Run single evaluation on combined files
-            overall_results = execute_evaluation(all_gold_path, all_pred_path)
-
-            outfile.write(f"Overall scores across all time periods ({len(valid_samples)} samples with a total of {total_sents} sentences):\n")
-            outfile.write("Metric\tPrecision\tRecall\tF1\tAligned Accuracy\n")
-
-            if overall_results:
-                for metric, score in overall_results.items():
-                    if metric in metrics:
-                        outfile.write(f"{metric}\t{score.precision:.4f}\t{score.recall:.4f}\t{score.f1:.4f}\t")
-                        if score.aligned_accuracy is None:
-                            outfile.write(f"{score.aligned_accuracy}\n")
-                        else:
-                            outfile.write(f"{score.aligned_accuracy:.4f}\n")
-            else:
-                outfile.write("Error computing overall scores.\n")
-
-    print(f"Results saved to {output_file}")
-
-
-# Run evaluation for all model configurations
-if __name__ == "__main__":
-    for predicted_dir, output_file in MODEL_CONFIGS:
-        run_evaluation(predicted_dir, output_file)
-    print("\nAll evaluations complete!")

eval/mod_conll18_ud_eval.py

@@ -1,672 +0,0 @@
-# This script is from the CoNLL 2018 Shared Task.
-# It was downloaded from here: https://universaldependencies.org/conll18/evaluation.html
-# in October 2025, and modified by Astrid B. Ingelstam for the group project about diachronic treebanks.
-
-#!/usr/bin/env python3
-
-# Compatible with Python 2.7 and 3.2+, can be used either as a module
-# or a standalone executable.
-#
-# Copyright 2017, 2018 Institute of Formal and Applied Linguistics (UFAL),
-# Faculty of Mathematics and Physics, Charles University, Czech Republic.
-#
-# This Source Code Form is subject to the terms of the Mozilla Public
-# License, v. 2.0. If a copy of the MPL was not distributed with this
-# file, You can obtain one at http://mozilla.org/MPL/2.0/.
-#
-# Authors: Milan Straka, Martin Popel <surname@ufal.mff.cuni.cz>
-#
-# Changelog:
-# - [12 Apr 2018] Version 0.9: Initial release.
-# - [19 Apr 2018] Version 1.0: Fix bug in MLAS (duplicate entries in functional_children).
-#                              Add --counts option.
-# - [02 May 2018] Version 1.1: When removing spaces to match gold and system characters,
-#                              consider all Unicode characters of category Zs instead of
-#                              just ASCII space.
-# - [25 Jun 2018] Version 1.2: Use python3 in the she-bang (instead of python).
-#                              In Python2, make the whole computation use `unicode` strings.
-
-# Command line usage
-# ------------------
-# conll18_ud_eval.py [-v] gold_conllu_file system_conllu_file
-#
-# - if no -v is given, only the official CoNLL18 UD Shared Task evaluation metrics
-#   are printed
-# - if -v is given, more metrics are printed (as precision, recall, F1 score,
-#   and in case the metric is computed on aligned words also accuracy on these):
-#   - Tokens: how well do the gold tokens match system tokens
-#   - Sentences: how well do the gold sentences match system sentences
-#   - Words: how well can the gold words be aligned to system words
-#   - UPOS: using aligned words, how well does UPOS match
-#   - XPOS: using aligned words, how well does XPOS match
-#   - UFeats: using aligned words, how well does universal FEATS match
-#   - AllTags: using aligned words, how well does UPOS+XPOS+FEATS match
-#   - Lemmas: using aligned words, how well does LEMMA match
-#   - UAS: using aligned words, how well does HEAD match
-#   - LAS: using aligned words, how well does HEAD+DEPREL(ignoring subtypes) match
-#   - CLAS: using aligned words with content DEPREL, how well does
-#       HEAD+DEPREL(ignoring subtypes) match
-#   - MLAS: using aligned words with content DEPREL, how well does
-#       HEAD+DEPREL(ignoring subtypes)+UPOS+UFEATS+FunctionalChildren(DEPREL+UPOS+UFEATS) match
-#   - BLEX: using aligned words with content DEPREL, how well does
-#       HEAD+DEPREL(ignoring subtypes)+LEMMAS match
-# - if -c is given, raw counts of correct/gold_total/system_total/aligned words are printed
-#   instead of precision/recall/F1/AlignedAccuracy for all metrics.
-
-# API usage
-# ---------
-# - load_conllu(file)
-#   - loads CoNLL-U file from given file object to an internal representation
-#   - the file object should return str in both Python 2 and Python 3
-#   - raises UDError exception if the given file cannot be loaded
-# - evaluate(gold_ud, system_ud)
-#   - evaluate the given gold and system CoNLL-U files (loaded with load_conllu)
-#   - raises UDError if the concatenated tokens of gold and system file do not match
-#   - returns a dictionary with the metrics described above, each metric having
-#     three fields: precision, recall and f1
-
-# Description of token matching
-# -----------------------------
-# In order to match tokens of gold file and system file, we consider the text
-# resulting from concatenation of gold tokens and text resulting from
-# concatenation of system tokens. These texts should match -- if they do not,
-# the evaluation fails.
-#
-# If the texts do match, every token is represented as a range in this original
-# text, and tokens are equal only if their range is the same.
-
-# Description of word matching
-# ----------------------------
-# When matching words of gold file and system file, we first match the tokens.
-# The words which are also tokens are matched as tokens, but words in multi-word
-# tokens have to be handled differently.
-#
-# To handle multi-word tokens, we start by finding "multi-word spans".
-# Multi-word span is a span in the original text such that
-# - it contains at least one multi-word token
-# - all multi-word tokens in the span (considering both gold and system ones)
-#   are completely inside the span (i.e., they do not "stick out")
-# - the multi-word span is as small as possible
-#
-# For every multi-word span, we align the gold and system words completely
-# inside this span using LCS on their FORMs. The words not intersecting
-# (even partially) any multi-word span are then aligned as tokens.
-
-
-from __future__ import division
-from __future__ import print_function
-
-import argparse
-import io
-import sys
-import unicodedata
-import unittest
-from collections import defaultdict
-
-CONFUSIONS = defaultdict(lambda: defaultdict(int)) #ADDED
-
-# CoNLL-U column names
-ID, FORM, LEMMA, UPOS, XPOS, FEATS, HEAD, DEPREL, DEPS, MISC = range(10)
-
-# Content and functional relations
-CONTENT_DEPRELS = {
-    "acl", "advcl", "advmod", "amod", "appos", "aux", "case", "cc", "ccomp",
-    "clf", "compound", "conj", "cop", "csubj", "dep", "det", "discourse",
-    "dislocated", "expl", "fixed", "flat", "goeswith", "iobj", "list", "mark",
-    "nmod", "nsubj", "nummod", "obj", "obl", "orphan", "parataxis", "punct",
-    "reparandum", "root", "vocative", "xcomp"
-}
-# Add for detailed scores
-UPOS_TAGS = {
-    "ADJ", "ADP", "ADV", "AUX", "CCONJ", "DET", "INTJ", "NOUN", "NUM", 
-    "PART", "PRON", "PROPN", "PUNCT", "SCONJ", "SYM", "VERB", "X"
-}
-##
-FUNCTIONAL_DEPRELS = {
-    "aux", "cop", "mark", "det", "clf", "case", "cc"
-}
-
-UNIVERSAL_FEATURES = {
-    "PronType", "NumType", "Poss", "Reflex", "Foreign", "Abbr", "Gender",
-    "Animacy", "Number", "Case", "Definite", "Degree", "VerbForm", "Mood",
-    "Tense", "Aspect", "Voice", "Evident", "Polarity", "Person", "Polite"
-}
-
-# UD Error is used when raising exceptions in this module
-class UDError(Exception):
-    pass
-
-# Conversion methods handling `str` <-> `unicode` conversions in Python2
-def _decode(text):
-    return text if sys.version_info[0] >= 3 or not isinstance(text, str) else text.decode("utf-8")
-
-def _encode(text):
-    return text if sys.version_info[0] >= 3 or not isinstance(text, unicode) else text.encode("utf-8")
-
-# Load given CoNLL-U file into internal representation
-def load_conllu(file, ignore_invalid_format=False):
-    # Internal representation classes
-    class UDRepresentation:
-        def __init__(self):
-            # Characters of all the tokens in the whole file.
-            # Whitespace between tokens is not included.
-            self.characters = []
-            # List of UDSpan instances with start&end indices into `characters`.
-            self.tokens = []
-            # List of UDWord instances.
-            self.words = []
-            # List of UDSpan instances with start&end indices into `characters`.
-            self.sentences = []
-    class UDSpan:
-        def __init__(self, start, end):
-            self.start = start
-            # Note that self.end marks the first position **after the end** of span,
-            # so we can use characters[start:end] or range(start, end).
-            self.end = end
-    class UDWord:
-        def __init__(self, span, columns, is_multiword):
-            # Span of this word (or MWT, see below) within ud_representation.characters.
-            self.span = span
-            # 10 columns of the CoNLL-U file: ID, FORM, LEMMA,...
-            self.columns = columns
-            # is_multiword==True means that this word is part of a multi-word token.
-            # In that case, self.span marks the span of the whole multi-word token.
-            self.is_multiword = is_multiword
-            # Reference to the UDWord instance representing the HEAD (or None if root).
-            self.parent = None
-            # List of references to UDWord instances representing functional-deprel children.
-            self.functional_children = []
-            # Only consider universal FEATS.
-            self.columns[FEATS] = "|".join(sorted(feat for feat in columns[FEATS].split("|")
-                                                  if feat.split("=", 1)[0] in UNIVERSAL_FEATURES))
-            # Let's ignore language-specific deprel subtypes.
-            self.columns[DEPREL] = columns[DEPREL].split(":")[0]
-            # Precompute which deprels are CONTENT_DEPRELS and which FUNCTIONAL_DEPRELS
-            self.is_content_deprel = self.columns[DEPREL] in CONTENT_DEPRELS
-            self.is_functional_deprel = self.columns[DEPREL] in FUNCTIONAL_DEPRELS
-
-    ud = UDRepresentation()
-
-    # Load the CoNLL-U file
-    index, sentence_start = 0, None
-    while True:
-        line = file.readline()
-        if not line:
-            break
-        line = _decode(line.rstrip("\r\n"))
-
-        # Handle sentence start boundaries
-        if sentence_start is None:
-            # Skip comments
-            if line.startswith("#"):
-                continue
-            # Start a new sentence
-            ud.sentences.append(UDSpan(index, 0))
-            sentence_start = len(ud.words)
-        if not line:
-            # Add parent and children UDWord links and check there are no cycles
-            def process_word(word):
-                if word.parent == "remapping":
-                    if not ignore_invalid_format:
-                        raise UDError("There is a cycle in a sentence")
-                if word.parent is None:
-                    try:
-                        head = int(word.columns[HEAD])
-                    except ValueError:
-                        head = 0 # Astrid: handle headless nodes as roots
-                    if head < 0 or head > len(ud.words) - sentence_start:
-                        if not ignore_invalid_format:
-                            raise UDError("HEAD '{}' points outside of the sentence".format(_encode(word.columns[HEAD])))
-                    if head:
-                        parent = ud.words[sentence_start + head - 1]
-                        word.parent = "remapping"
-                        process_word(parent)
-                        word.parent = parent
-
-            for word in ud.words[sentence_start:]:
-                process_word(word)
-            # func_children cannot be assigned within process_word
-            # because it is called recursively and may result in adding one child twice.
-            for word in ud.words[sentence_start:]:
-                if word.parent and word.is_functional_deprel:
-                    word.parent.functional_children.append(word)
-
-            # Check there is a single root node
-            if len([word for word in ud.words[sentence_start:] if word.parent is None]) != 1:
-                #print(f"Warning: There are {len([word for word in ud.words[sentence_start:] if word.parent is None])} roots in a sentence.")
-                if not ignore_invalid_format:
-                    raise UDError("There are multiple roots in a sentence")
-
-            # End the sentence
-            ud.sentences[-1].end = index
-            sentence_start = None
-            continue
-
-        # Read next token/word
-        columns = line.split("\t")
-        if len(columns) != 10:
-            raise UDError("The CoNLL-U line does not contain 10 tab-separated columns: '{}'".format(_encode(line)))
-
-        # Skip empty nodes
-        if "." in columns[ID]:
-            continue
-
-        # Delete spaces from FORM, so gold.characters == system.characters
-        # even if one of them tokenizes the space. Use any Unicode character
-        # with category Zs.
-        columns[FORM] = "".join(filter(lambda c: unicodedata.category(c) != "Zs", columns[FORM]))
-        if not columns[FORM]:
-            raise UDError("There is an empty FORM in the CoNLL-U file")
-
-        # Save token
-        ud.characters.extend(columns[FORM])
-        ud.tokens.append(UDSpan(index, index + len(columns[FORM])))
-        index += len(columns[FORM])
-
-        # Handle multi-word tokens to save word(s)
-        if "-" in columns[ID]:
-            try:
-                start, end = map(int, columns[ID].split("-"))
-            except:
-                raise UDError("Cannot parse multi-word token ID '{}'".format(_encode(columns[ID])))
-
-            for _ in range(start, end + 1):
-                word_line = _decode(file.readline().rstrip("\r\n"))
-                word_columns = word_line.split("\t")
-                if len(word_columns) != 10:
-                    raise UDError("The CoNLL-U line does not contain 10 tab-separated columns: '{}'".format(_encode(word_line)))
-                ud.words.append(UDWord(ud.tokens[-1], word_columns, is_multiword=True))
-        # Basic tokens/words
-        else:
-            try:
-                word_id = int(columns[ID])
-            except:
-                raise UDError("Cannot parse word ID '{}'".format(_encode(columns[ID])))
-            if word_id != len(ud.words) - sentence_start + 1:
-                raise UDError("Incorrect word ID '{}' for word '{}', expected '{}'".format(
-                    _encode(columns[ID]), _encode(columns[FORM]), len(ud.words) - sentence_start + 1))
-
-            try:
-                head_id = int(columns[HEAD])
-            except:
-                if ignore_invalid_format:
-                    head_id = 0 # Astrid: handle headless nodes as roots
-                else:
-                    raise UDError("Cannot parse HEAD '{}'".format(_encode(columns[HEAD])))
-            if head_id < 0:
-                raise UDError("HEAD cannot be negative")
-
-            ud.words.append(UDWord(ud.tokens[-1], columns, is_multiword=False))
-
-    if sentence_start is not None:
-        raise UDError("The CoNLL-U file does not end with empty line")
-
-    return ud
-
-# Evaluate the gold and system treebanks (loaded using load_conllu).
-def evaluate(gold_ud, system_ud, detailed=False, cm=False):  #ADDED cm=False
-    class Score:
-        def __init__(self, gold_total, system_total, correct, aligned_total=None):
-            self.correct = correct
-            self.gold_total = gold_total
-            self.system_total = system_total
-            self.aligned_total = aligned_total
-            self.precision = correct / system_total if system_total else 0.0
-            self.recall = correct / gold_total if gold_total else 0.0
-            self.f1 = 2 * correct / (system_total + gold_total) if system_total + gold_total else 0.0
-            self.aligned_accuracy = correct / aligned_total if aligned_total else aligned_total
-    class AlignmentWord:
-        def __init__(self, gold_word, system_word):
-            self.gold_word = gold_word
-            self.system_word = system_word
-    class Alignment:
-        def __init__(self, gold_words, system_words):
-            self.gold_words = gold_words
-            self.system_words = system_words
-            self.matched_words = []
-            self.matched_words_map = {}
-        def append_aligned_words(self, gold_word, system_word):
-            self.matched_words.append(AlignmentWord(gold_word, system_word))
-            self.matched_words_map[system_word] = gold_word
-
-    def spans_score(gold_spans, system_spans):
-        correct, gi, si = 0, 0, 0
-        while gi < len(gold_spans) and si < len(system_spans):
-            if system_spans[si].start < gold_spans[gi].start:
-                si += 1
-            elif gold_spans[gi].start < system_spans[si].start:
-                gi += 1
-            else:
-                correct += gold_spans[gi].end == system_spans[si].end
-                si += 1
-                gi += 1
-
-        return Score(len(gold_spans), len(system_spans), correct)
-
-    def alignment_score(alignment, key_fn=None, filter_fn=None, name=None): # added "name" flag for confusion matrix
-        if filter_fn is not None:
-            gold = sum(1 for gold in alignment.gold_words if filter_fn(gold))
-            system = sum(1 for system in alignment.system_words if filter_fn(system))
-            aligned = sum(
-                1 for pair in alignment.matched_words 
-                if filter_fn(pair.gold_word) and filter_fn(pair.system_word))
-        else:
-            gold = len(alignment.gold_words)
-            system = len(alignment.system_words)
-            aligned = len(alignment.matched_words)
-
-        if key_fn is None:
-            # Return score for whole aligned words
-            return Score(gold, system, aligned)
-
-        def gold_aligned_gold(word):
-            return word
-        def gold_aligned_system(word):
-            return alignment.matched_words_map.get(word, "NotAligned") if word is not None else None
-        correct = 0
-        for words in alignment.matched_words:
-            if filter_fn is None or filter_fn(words.gold_word):
-                gold_label = key_fn(words.gold_word, gold_aligned_gold)    # ADDED
-                system_label = key_fn(words.system_word, gold_aligned_system)  # ADDED
-                if key_fn(words.gold_word, gold_aligned_gold) == key_fn(words.system_word, gold_aligned_system):
-                    correct += 1
-
-                if name is not None:
-                    CONFUSIONS[name][(gold_label, system_label)] += 1
-
-        return Score(gold, system, correct, aligned)
-
-    def beyond_end(words, i, multiword_span_end):
-        if i >= len(words):
-            return True
-        if words[i].is_multiword:
-            return words[i].span.start >= multiword_span_end
-        return words[i].span.end > multiword_span_end
-
-    def extend_end(word, multiword_span_end):
-        if word.is_multiword and word.span.end > multiword_span_end:
-            return word.span.end
-        return multiword_span_end
-
-    def find_multiword_span(gold_words, system_words, gi, si):
-        # We know gold_words[gi].is_multiword or system_words[si].is_multiword.
-        # Find the start of the multiword span (gs, ss), so the multiword span is minimal.
-        # Initialize multiword_span_end characters index.
-        if gold_words[gi].is_multiword:
-            multiword_span_end = gold_words[gi].span.end
-            if not system_words[si].is_multiword and system_words[si].span.start < gold_words[gi].span.start:
-                si += 1
-        else: # if system_words[si].is_multiword
-            multiword_span_end = system_words[si].span.end
-            if not gold_words[gi].is_multiword and gold_words[gi].span.start < system_words[si].span.start:
-                gi += 1
-        gs, ss = gi, si
-
-        # Find the end of the multiword span
-        # (so both gi and si are pointing to the word following the multiword span end).
-        while not beyond_end(gold_words, gi, multiword_span_end) or \
-              not beyond_end(system_words, si, multiword_span_end):
-            if gi < len(gold_words) and (si >= len(system_words) or
-                                         gold_words[gi].span.start <= system_words[si].span.start):
-                multiword_span_end = extend_end(gold_words[gi], multiword_span_end)
-                gi += 1
-            else:
-                multiword_span_end = extend_end(system_words[si], multiword_span_end)
-                si += 1
-        return gs, ss, gi, si
-
-    def compute_lcs(gold_words, system_words, gi, si, gs, ss):
-        lcs = [[0] * (si - ss) for i in range(gi - gs)]
-        for g in reversed(range(gi - gs)):
-            for s in reversed(range(si - ss)):
-                if gold_words[gs + g].columns[FORM].lower() == system_words[ss + s].columns[FORM].lower():
-                    lcs[g][s] = 1 + (lcs[g+1][s+1] if g+1 < gi-gs and s+1 < si-ss else 0)
-                lcs[g][s] = max(lcs[g][s], lcs[g+1][s] if g+1 < gi-gs else 0)
-                lcs[g][s] = max(lcs[g][s], lcs[g][s+1] if s+1 < si-ss else 0)
-        return lcs
-
-    def align_words(gold_words, system_words):
-        alignment = Alignment(gold_words, system_words)
-
-        gi, si = 0, 0
-        while gi < len(gold_words) and si < len(system_words):
-            if gold_words[gi].is_multiword or system_words[si].is_multiword:
-                # A: Multi-word tokens => align via LCS within the whole "multiword span".
-                gs, ss, gi, si = find_multiword_span(gold_words, system_words, gi, si)
-
-                if si > ss and gi > gs:
-                    lcs = compute_lcs(gold_words, system_words, gi, si, gs, ss)
-
-                    # Store aligned words
-                    s, g = 0, 0
-                    while g < gi - gs and s < si - ss:
-                        if gold_words[gs + g].columns[FORM].lower() == system_words[ss + s].columns[FORM].lower():
-                            alignment.append_aligned_words(gold_words[gs+g], system_words[ss+s])
-                            g += 1
-                            s += 1
-                        elif lcs[g][s] == (lcs[g+1][s] if g+1 < gi-gs else 0):
-                            g += 1
-                        else:
-                            s += 1
-            else:
-                # B: No multi-word token => align according to spans.
-                if (gold_words[gi].span.start, gold_words[gi].span.end) == (system_words[si].span.start, system_words[si].span.end):
-                    alignment.append_aligned_words(gold_words[gi], system_words[si])
-                    gi += 1
-                    si += 1
-                elif gold_words[gi].span.start <= system_words[si].span.start:
-                    gi += 1
-                else:
-                    si += 1
-
-        return alignment
-
-    # Check that the underlying character sequences do match.
-    if gold_ud.characters != system_ud.characters:
-        """index = 0
-        while index < len(gold_ud.characters) and index < len(system_ud.characters) and \
-                gold_ud.characters[index] == system_ud.characters[index]:
-            index += 1
-
-        raise UDError(
-            "The concatenation of tokens in gold file and in system file differ!\n" +
-            "First 20 differing characters in gold file: '{}' and system file: '{}'".format(
-                "".join(map(_encode, gold_ud.characters[index:index + 20])),
-                "".join(map(_encode, system_ud.characters[index:index + 20]))
-            )
-        )"""
-        gold_ud.characters = system_ud.characters # Astrid: token form is altered - go with base tree form
-
-
-    # Added to compute detailed UPOS scores:
-    def detailed_score(alignment, labels, column_ind):
-        r = {}
-        total_gold, total_system, total_correct = 0, 0, 0
-        for label in labels:
-            score = alignment_score(
-                alignment,
-                filter_fn=lambda w, label=label: w.columns[column_ind] == label
-                )
-
-            r[label] = {
-                "precision": score.precision,
-                "recall": score.recall,
-                "f1": score.f1
-            }
-            total_gold += score.gold_total
-            total_system += score.system_total
-            total_correct += score.correct
-
-        r[f'TOTAL'] = {
-            "precision": total_correct / total_system if total_system else 0.0,
-            "recall": total_correct / total_gold if total_gold else 0.0,
-            "f1": 2 * total_correct / (total_system + total_gold) if (total_system + total_gold) else 0.0
-        }
-        return r
-    
-    # Added to compute detailed UPOS scores:
-
-    ## ADDED
-    def dump_confusion(confusion, filename):
-        labels = sorted(set(g for g, s in confusion) | set(s for g, s in confusion))
-        with open(filename, "w", encoding="utf-8") as f:
-            f.write("gold\\system\t" + "\t".join(map(str, labels)) + "\n")
-            for g in labels:
-                row = [str(confusion.get((g, s), 0)) for s in labels]
-                f.write(str(g) + "\t" + "\t".join(row) + "\n")
-        
-    #for name, matrix in CONFUSIONS.items():
-       # dump_confusion(matrix, f"{name}_confusion.tsv")
-    ## ADDED
-
-
-    # Align words
-    alignment = align_words(gold_ud.words, system_ud.words)
-
-    # Compute the F1-scores
-    if detailed: #The initial F1-scores
-        upos_scores = detailed_score(alignment, UPOS_TAGS, UPOS)
-        deprel_scores = detailed_score(alignment, CONTENT_DEPRELS, DEPREL)
-        return upos_scores, deprel_scores
-    
-    # For Confusion matrix
-    if cm:
-        sc = {
-            "UPOS": alignment_score(alignment,lambda w, _: w.columns[UPOS], name = "UPOS"),
-            "DEPREL": alignment_score(alignment, lambda w, _: w.columns[DEPREL], name ="DEPREL"),
-        }
-        return CONFUSIONS
-    
-
-    return {
-        "Tokens": spans_score(gold_ud.tokens, system_ud.tokens),
-        "Sentences": spans_score(gold_ud.sentences, system_ud.sentences),
-        "Words": alignment_score(alignment),
-        "UPOS": alignment_score(alignment, lambda w, _: w.columns[UPOS]),
-        "XPOS": alignment_score(alignment, lambda w, _: w.columns[XPOS]),
-        "UFeats": alignment_score(alignment, lambda w, _: w.columns[FEATS]),
-        "AllTags": alignment_score(alignment, lambda w, _: (w.columns[UPOS], w.columns[XPOS], w.columns[FEATS])),
-        "Lemmas": alignment_score(alignment, lambda w, ga: w.columns[LEMMA] if ga(w).columns[LEMMA] != "_" else "_"),
-        "UAS": alignment_score(alignment, lambda w, ga: ga(w.parent)),
-        "LAS": alignment_score(alignment, lambda w, ga: (ga(w.parent), w.columns[DEPREL])),
-        "CLAS": alignment_score(alignment, lambda w, ga: (ga(w.parent), w.columns[DEPREL]),
-                                filter_fn=lambda w: w.is_content_deprel),
-        "MLAS": alignment_score(alignment, lambda w, ga: (ga(w.parent), w.columns[DEPREL], w.columns[UPOS], w.columns[FEATS],
-                                                        [(ga(c), c.columns[DEPREL], c.columns[UPOS], c.columns[FEATS])
-                                                        for c in w.functional_children]),
-                                filter_fn=lambda w: w.is_content_deprel),
-        "BLEX": alignment_score(alignment, lambda w, ga: (ga(w.parent), w.columns[DEPREL],
-                                                        w.columns[LEMMA] if ga(w).columns[LEMMA] != "_" else "_"),
-                                filter_fn=lambda w: w.is_content_deprel),
-    }
-   
-
-def load_conllu_file(path, ignore_invalid_format=False):
-    _file = open(path, mode="r", **({"encoding": "utf-8"} if sys.version_info >= (3, 0) else {}))
-    return load_conllu(_file, ignore_invalid_format=ignore_invalid_format)
-
-def evaluate_wrapper(args):
-    # Load CoNLL-U files
-    gold_ud = load_conllu_file(args.gold_file)
-    system_ud = load_conllu_file(args.system_file, ignore_invalid_format=args.ignore_invalid_format)
-    return evaluate(gold_ud, system_ud)
-
-def main():
-    # Parse arguments
-    parser = argparse.ArgumentParser()
-    parser.add_argument("gold_file", type=str,
-                        help="Name of the CoNLL-U file with the gold data.")
-    parser.add_argument("system_file", type=str,
-                        help="Name of the CoNLL-U file with the predicted data.")
-    parser.add_argument("--verbose", "-v", default=False, action="store_true",
-                        help="Print all metrics.")
-    parser.add_argument("--counts", "-c", default=False, action="store_true",
-                        help="Print raw counts of correct/gold/system/aligned words instead of prec/rec/F1 for all metrics.")
-    parser.add_argument("--ignore-invalid-format", "-i", default=False, action="store_true",
-                        help="Ignore invalid format errors, such as multiple roots, when loading CoNLL-U files.")
-    args = parser.parse_args()
-
-    # Evaluate
-    evaluation = evaluate_wrapper(args)
-
-    # Print the evaluation
-    if not args.verbose and not args.counts:
-        print("LAS F1 Score: {:.2f}".format(100 * evaluation["LAS"].f1))
-        print("MLAS Score: {:.2f}".format(100 * evaluation["MLAS"].f1))
-        print("BLEX Score: {:.2f}".format(100 * evaluation["BLEX"].f1))
-    else:
-        if args.counts:
-            print("Metric     | Correct   |      Gold | Predicted | Aligned")
-        else:
-            print("Metric     | Precision |    Recall |  F1 Score | AligndAcc")
-        print("-----------+-----------+-----------+-----------+-----------")
-        for metric in["Tokens", "Sentences", "Words", "UPOS", "XPOS", "UFeats", "AllTags", "Lemmas", "UAS", "LAS", "CLAS", "MLAS", "BLEX"]:
-            if args.counts:
-                print("{:11}|{:10} |{:10} |{:10} |{:10}".format(
-                    metric,
-                    evaluation[metric].correct,
-                    evaluation[metric].gold_total,
-                    evaluation[metric].system_total,
-                    evaluation[metric].aligned_total or (evaluation[metric].correct if metric == "Words" else "")
-                ))
-            else:
-                print("{:11}|{:10.2f} |{:10.2f} |{:10.2f} |{}".format(
-                    metric,
-                    100 * evaluation[metric].precision,
-                    100 * evaluation[metric].recall,
-                    100 * evaluation[metric].f1,
-                    "{:10.2f}".format(100 * evaluation[metric].aligned_accuracy) if evaluation[metric].aligned_accuracy is not None else ""
-                ))
-
-if __name__ == "__main__":
-    main()
-
-# Tests, which can be executed with `python -m unittest conll18_ud_eval`.
-class TestAlignment(unittest.TestCase):
-    @staticmethod
-    def _load_words(words):
-        """Prepare fake CoNLL-U files with fake HEAD to prevent multiple roots errors."""
-        lines, num_words = [], 0
-        for w in words:
-            parts = w.split(" ")
-            if len(parts) == 1:
-                num_words += 1
-                lines.append("{}\t{}\t_\t_\t_\t_\t{}\t_\t_\t_".format(num_words, parts[0], int(num_words>1)))
-            else:
-                lines.append("{}-{}\t{}\t_\t_\t_\t_\t_\t_\t_\t_".format(num_words + 1, num_words + len(parts) - 1, parts[0]))
-                for part in parts[1:]:
-                    num_words += 1
-                    lines.append("{}\t{}\t_\t_\t_\t_\t{}\t_\t_\t_".format(num_words, part, int(num_words>1)))
-        return load_conllu((io.StringIO if sys.version_info >= (3, 0) else io.BytesIO)("\n".join(lines+["\n"])))
-
-    def _test_exception(self, gold, system):
-        self.assertRaises(UDError, evaluate, self._load_words(gold), self._load_words(system))
-
-    def _test_ok(self, gold, system, correct):
-        metrics = evaluate(self._load_words(gold), self._load_words(system))
-        gold_words = sum((max(1, len(word.split(" ")) - 1) for word in gold))
-        system_words = sum((max(1, len(word.split(" ")) - 1) for word in system))
-        self.assertEqual((metrics["Words"].precision, metrics["Words"].recall, metrics["Words"].f1),
-                         (correct / system_words, correct / gold_words, 2 * correct / (gold_words + system_words)))
-
-    def test_exception(self):
-        self._test_exception(["a"], ["b"])
-
-    def test_equal(self):
-        self._test_ok(["a"], ["a"], 1)
-        self._test_ok(["a", "b", "c"], ["a", "b", "c"], 3)
-
-    def test_equal_with_multiword(self):
-        self._test_ok(["abc a b c"], ["a", "b", "c"], 3)
-        self._test_ok(["a", "bc b c", "d"], ["a", "b", "c", "d"], 4)
-        self._test_ok(["abcd a b c d"], ["ab a b", "cd c d"], 4)
-        self._test_ok(["abc a b c", "de d e"], ["a", "bcd b c d", "e"], 5)
-
-    def test_alignment(self):
-        self._test_ok(["abcd"], ["a", "b", "c", "d"], 0)
-        self._test_ok(["abc", "d"], ["a", "b", "c", "d"], 1)
-        self._test_ok(["a", "bc", "d"], ["a", "b", "c", "d"], 2)
-        self._test_ok(["a", "bc b c", "d"], ["a", "b", "cd"], 2)
-        self._test_ok(["abc a BX c", "def d EX f"], ["ab a b", "cd c d", "ef e f"], 4)
-        self._test_ok(["ab a b", "cd bc d"], ["a", "bc", "d"], 2)
-        self._test_ok(["a", "bc b c", "d"], ["ab AX BX", "cd CX a"], 1)

eval/scores/scores_talbanken.txt

@@ -1,162 +0,0 @@
-Sample: svediakorp-sec987-spf144, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	1.0000	1.0000	1.0000	1.0000
-UAS	1.0000	1.0000	1.0000	1.0000
-LAS	0.9615	0.9615	0.9615	0.9615
-
-Sample: svediakorp-sec613-EngstromA_StrindbergOchJag, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	1.0000	1.0000	1.0000	1.0000
-UAS	0.9146	0.9146	0.9146	0.9146
-LAS	0.8659	0.8659	0.8659	0.8659
-
-Sample: svediakorp-sec268-DulciU_VitterhetsNojen3, time period: 1750-1800
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9663	0.9663	0.9663	0.9663
-UAS	0.8798	0.8798	0.8798	0.8798
-LAS	0.8365	0.8365	0.8365	0.8365
-
-Sample: svediakorp-sec25-Runius, time period: 1700-1750
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.8750	0.8750	0.8750	0.8750
-UAS	0.7188	0.7188	0.7188	0.7188
-LAS	0.6562	0.6562	0.6562	0.6562
-
-Sample: svediakorp-sec208-Anonym_DetGrasligaMordet, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9865	0.9865	0.9865	0.9865
-UAS	0.8784	0.8784	0.8784	0.8784
-LAS	0.8243	0.8243	0.8243	0.8243
-
-Sample: svediakorp-sec324-GranbergPA_Enslighetsalskaren, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9400	0.9400	0.9400	0.9400
-UAS	0.8400	0.8400	0.8400	0.8400
-LAS	0.7400	0.7400	0.7400	0.7400
-
-Sample: svediakorp-sec254-CederborghF_BerattelseOmJohnHall, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9706	0.9706	0.9706	0.9706
-UAS	0.5412	0.5412	0.5412	0.5412
-LAS	0.5000	0.5000	0.5000	0.5000
-
-Sample: svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9800	0.9800	0.9800	0.9800
-UAS	0.7200	0.7200	0.7200	0.7200
-LAS	0.7200	0.7200	0.7200	0.7200
-
-Sample: svediakorp-sec452-NyblomH_FantasierFyra, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9464	0.9464	0.9464	0.9464
-UAS	0.7500	0.7500	0.7500	0.7500
-LAS	0.6786	0.6786	0.6786	0.6786
-
-Sample: svediakorp-sec252-BremerF_Teckningar1, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	1.0000	1.0000	1.0000	1.0000
-UAS	0.8750	0.8750	0.8750	0.8750
-LAS	0.8750	0.8750	0.8750	0.8750
-
-Sample: svediakorp-sec486-SchwartzMS_BellmansSkor, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9873	0.9873	0.9873	0.9873
-UAS	0.7722	0.7722	0.7722	0.7722
-LAS	0.6962	0.6962	0.6962	0.6962
-
-Sample: svediakorp-sec330-GyllenborgC_SwenskaSpratthoken, time period: 1700-1750
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.8650	0.8650	0.8650	0.8650
-UAS	0.6687	0.6687	0.6687	0.6687
-LAS	0.5828	0.5828	0.5828	0.5828
-
-Sample: svediakorp-sec631-HasselskogN_HallaHallaGronkoping, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9605	0.9605	0.9605	0.9605
-UAS	0.7368	0.7368	0.7368	0.7368
-LAS	0.6974	0.6974	0.6974	0.6974
-
-Sample: svediakorp-sec277-EnbomPU_MedborgeligtSkalde, time period: 1750-1800
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9464	0.9464	0.9464	0.9464
-UAS	0.7500	0.7500	0.7500	0.7500
-LAS	0.6786	0.6786	0.6786	0.6786
-
-Sample: svediakorp-sec1102-spf259, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9444	0.9444	0.9444	0.9444
-UAS	0.8889	0.8889	0.8889	0.8889
-LAS	0.8889	0.8889	0.8889	0.8889
-
-Sample: svediakorp-sec991-spf148, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9561	0.9561	0.9561	0.9561
-UAS	0.7982	0.7982	0.7982	0.7982
-LAS	0.7368	0.7368	0.7368	0.7368
-
-Sample: svediakorp-sec639-HeidenstamV_Proletarfilosofiens, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9861	0.9861	0.9861	0.9861
-UAS	1.0000	1.0000	1.0000	1.0000
-LAS	1.0000	1.0000	1.0000	1.0000
-
-Sample: svediakorp-sec988-spf145, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9147	0.9147	0.9147	0.9147
-UAS	0.8450	0.8450	0.8450	0.8450
-LAS	0.7829	0.7829	0.7829	0.7829
-
-Sample: svediakorp-sec1063-spf220, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9660	0.9660	0.9660	0.9660
-UAS	0.9184	0.9184	0.9184	0.9184
-LAS	0.8707	0.8707	0.8707	0.8707
-
-Sample: svediakorp-sec1033-spf190, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9692	0.9692	0.9692	0.9692
-UAS	0.8923	0.8923	0.8923	0.8923
-LAS	0.8462	0.8462	0.8462	0.8462
-
-Sample: svediakorp-letter141673-Stalhammar, time period: 1700-1750
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.7059	0.7059	0.7059	0.7059
-UAS	0.5294	0.5294	0.5294	0.5294
-LAS	0.4118	0.4118	0.4118	0.4118
-
-----------------------------------------------------------------------
-Average scores for time period 1700-1750 (3 samples with a total of 8 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.8302	0.8302	0.8302	0.8302
-UAS	0.6589	0.6589	0.6589	0.6589
-LAS	0.5768	0.5768	0.5768	0.5768
-
-Average scores for time period 1750-1800 (2 samples with a total of 5 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9584	0.9584	0.9584	0.9584
-UAS	0.8279	0.8279	0.8279	0.8279
-LAS	0.7734	0.7734	0.7734	0.7734
-
-Average scores for time period 1800-1850 (6 samples with a total of 19 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9517	0.9517	0.9517	0.9517
-UAS	0.7774	0.7774	0.7774	0.7774
-LAS	0.7164	0.7164	0.7164	0.7164
-
-Average scores for time period 1850-1900 (5 samples with a total of 14 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9692	0.9692	0.9692	0.9692
-UAS	0.8552	0.8552	0.8552	0.8552
-LAS	0.8006	0.8006	0.8006	0.8006
-
-Average scores for time period 1900-1950 (5 samples with a total of 13 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9805	0.9805	0.9805	0.9805
-UAS	0.8719	0.8719	0.8719	0.8719
-LAS	0.8439	0.8439	0.8439	0.8439
-
-Overall scores across all time periods (21 samples with a total of 59 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9511	0.9511	0.9511	0.9511
-UAS	0.7992	0.7992	0.7992	0.7992
-LAS	0.7452	0.7452	0.7452	0.7452

eval/scores/scores_transformer_no_silver.txt

@@ -1,162 +0,0 @@
-Sample: svediakorp-sec987-spf144, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	1.0000	1.0000	1.0000	1.0000
-UAS	0.9615	0.9615	0.9615	0.9615
-LAS	0.9231	0.9231	0.9231	0.9231
-
-Sample: svediakorp-sec613-EngstromA_StrindbergOchJag, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	1.0000	1.0000	1.0000	1.0000
-UAS	0.9024	0.9024	0.9024	0.9024
-LAS	0.8537	0.8537	0.8537	0.8537
-
-Sample: svediakorp-sec268-DulciU_VitterhetsNojen3, time period: 1750-1800
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9663	0.9663	0.9663	0.9663
-UAS	0.8317	0.8317	0.8317	0.8317
-LAS	0.7837	0.7837	0.7837	0.7837
-
-Sample: svediakorp-sec25-Runius, time period: 1700-1750
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.8750	0.8750	0.8750	0.8750
-UAS	0.6875	0.6875	0.6875	0.6875
-LAS	0.5625	0.5625	0.5625	0.5625
-
-Sample: svediakorp-sec208-Anonym_DetGrasligaMordet, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9865	0.9865	0.9865	0.9865
-UAS	0.7973	0.7973	0.7973	0.7973
-LAS	0.7432	0.7432	0.7432	0.7432
-
-Sample: svediakorp-sec324-GranbergPA_Enslighetsalskaren, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9400	0.9400	0.9400	0.9400
-UAS	0.8800	0.8800	0.8800	0.8800
-LAS	0.8100	0.8100	0.8100	0.8100
-
-Sample: svediakorp-sec254-CederborghF_BerattelseOmJohnHall, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9706	0.9706	0.9706	0.9706
-UAS	0.5882	0.5882	0.5882	0.5882
-LAS	0.5118	0.5118	0.5118	0.5118
-
-Sample: svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9800	0.9800	0.9800	0.9800
-UAS	0.7800	0.7800	0.7800	0.7800
-LAS	0.7800	0.7800	0.7800	0.7800
-
-Sample: svediakorp-sec452-NyblomH_FantasierFyra, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9464	0.9464	0.9464	0.9464
-UAS	0.8393	0.8393	0.8393	0.8393
-LAS	0.8036	0.8036	0.8036	0.8036
-
-Sample: svediakorp-sec252-BremerF_Teckningar1, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	1.0000	1.0000	1.0000	1.0000
-UAS	0.8750	0.8750	0.8750	0.8750
-LAS	0.8750	0.8750	0.8750	0.8750
-
-Sample: svediakorp-sec486-SchwartzMS_BellmansSkor, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9873	0.9873	0.9873	0.9873
-UAS	0.7848	0.7848	0.7848	0.7848
-LAS	0.6835	0.6835	0.6835	0.6835
-
-Sample: svediakorp-sec330-GyllenborgC_SwenskaSpratthoken, time period: 1700-1750
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.8650	0.8650	0.8650	0.8650
-UAS	0.7301	0.7301	0.7301	0.7301
-LAS	0.6135	0.6135	0.6135	0.6135
-
-Sample: svediakorp-sec631-HasselskogN_HallaHallaGronkoping, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9605	0.9605	0.9605	0.9605
-UAS	0.8684	0.8684	0.8684	0.8684
-LAS	0.7632	0.7632	0.7632	0.7632
-
-Sample: svediakorp-sec277-EnbomPU_MedborgeligtSkalde, time period: 1750-1800
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9464	0.9464	0.9464	0.9464
-UAS	0.7857	0.7857	0.7857	0.7857
-LAS	0.7500	0.7500	0.7500	0.7500
-
-Sample: svediakorp-sec1102-spf259, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9444	0.9444	0.9444	0.9444
-UAS	0.8889	0.8889	0.8889	0.8889
-LAS	0.8889	0.8889	0.8889	0.8889
-
-Sample: svediakorp-sec991-spf148, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9561	0.9561	0.9561	0.9561
-UAS	0.7982	0.7982	0.7982	0.7982
-LAS	0.7807	0.7807	0.7807	0.7807
-
-Sample: svediakorp-sec639-HeidenstamV_Proletarfilosofiens, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9861	0.9861	0.9861	0.9861
-UAS	0.9583	0.9583	0.9583	0.9583
-LAS	0.9444	0.9444	0.9444	0.9444
-
-Sample: svediakorp-sec988-spf145, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9147	0.9147	0.9147	0.9147
-UAS	0.8682	0.8682	0.8682	0.8682
-LAS	0.7984	0.7984	0.7984	0.7984
-
-Sample: svediakorp-sec1063-spf220, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9660	0.9660	0.9660	0.9660
-UAS	0.9116	0.9116	0.9116	0.9116
-LAS	0.8503	0.8503	0.8503	0.8503
-
-Sample: svediakorp-sec1033-spf190, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9692	0.9692	0.9692	0.9692
-UAS	0.9077	0.9077	0.9077	0.9077
-LAS	0.8154	0.8154	0.8154	0.8154
-
-Sample: svediakorp-letter141673-Stalhammar, time period: 1700-1750
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.7059	0.7059	0.7059	0.7059
-UAS	0.6471	0.6471	0.6471	0.6471
-LAS	0.4706	0.4706	0.4706	0.4706
-
-----------------------------------------------------------------------
-Average scores for time period 1700-1750 (3 samples with a total of 8 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.8302	0.8302	0.8302	0.8302
-UAS	0.6880	0.6880	0.6880	0.6880
-LAS	0.5523	0.5523	0.5523	0.5523
-
-Average scores for time period 1750-1800 (2 samples with a total of 5 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9584	0.9584	0.9584	0.9584
-UAS	0.8133	0.8133	0.8133	0.8133
-LAS	0.7702	0.7702	0.7702	0.7702
-
-Average scores for time period 1800-1850 (6 samples with a total of 19 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9517	0.9517	0.9517	0.9517
-UAS	0.7986	0.7986	0.7986	0.7986
-LAS	0.7464	0.7464	0.7464	0.7464
-
-Average scores for time period 1850-1900 (5 samples with a total of 14 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9692	0.9692	0.9692	0.9692
-UAS	0.8561	0.8561	0.8561	0.8561
-LAS	0.7954	0.7954	0.7954	0.7954
-
-Average scores for time period 1900-1950 (5 samples with a total of 13 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9805	0.9805	0.9805	0.9805
-UAS	0.8925	0.8925	0.8925	0.8925
-LAS	0.8405	0.8405	0.8405	0.8405
-
-Overall scores across all time periods (21 samples with a total of 59 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9511	0.9511	0.9511	0.9511
-UAS	0.8140	0.8140	0.8140	0.8140
-LAS	0.7503	0.7503	0.7503	0.7503

eval/scores/scores_transformer_silver.txt

@@ -1,162 +0,0 @@
-Sample: svediakorp-sec987-spf144, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	1.0000	1.0000	1.0000	1.0000
-UAS	0.9615	0.9615	0.9615	0.9615
-LAS	0.8846	0.8846	0.8846	0.8846
-
-Sample: svediakorp-sec613-EngstromA_StrindbergOchJag, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	1.0000	1.0000	1.0000	1.0000
-UAS	0.8780	0.8780	0.8780	0.8780
-LAS	0.8171	0.8171	0.8171	0.8171
-
-Sample: svediakorp-sec268-DulciU_VitterhetsNojen3, time period: 1750-1800
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9663	0.9663	0.9663	0.9663
-UAS	0.8221	0.8221	0.8221	0.8221
-LAS	0.7692	0.7692	0.7692	0.7692
-
-Sample: svediakorp-sec25-Runius, time period: 1700-1750
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.8750	0.8750	0.8750	0.8750
-UAS	0.6562	0.6562	0.6562	0.6562
-LAS	0.5625	0.5625	0.5625	0.5625
-
-Sample: svediakorp-sec208-Anonym_DetGrasligaMordet, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9865	0.9865	0.9865	0.9865
-UAS	0.7568	0.7568	0.7568	0.7568
-LAS	0.7027	0.7027	0.7027	0.7027
-
-Sample: svediakorp-sec324-GranbergPA_Enslighetsalskaren, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9400	0.9400	0.9400	0.9400
-UAS	0.8500	0.8500	0.8500	0.8500
-LAS	0.7500	0.7500	0.7500	0.7500
-
-Sample: svediakorp-sec254-CederborghF_BerattelseOmJohnHall, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9706	0.9706	0.9706	0.9706
-UAS	0.5765	0.5765	0.5765	0.5765
-LAS	0.5000	0.5000	0.5000	0.5000
-
-Sample: svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9800	0.9800	0.9800	0.9800
-UAS	0.8000	0.8000	0.8000	0.8000
-LAS	0.7600	0.7600	0.7600	0.7600
-
-Sample: svediakorp-sec452-NyblomH_FantasierFyra, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9464	0.9464	0.9464	0.9464
-UAS	0.8571	0.8571	0.8571	0.8571
-LAS	0.7679	0.7679	0.7679	0.7679
-
-Sample: svediakorp-sec252-BremerF_Teckningar1, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	1.0000	1.0000	1.0000	1.0000
-UAS	0.8333	0.8333	0.8333	0.8333
-LAS	0.7917	0.7917	0.7917	0.7917
-
-Sample: svediakorp-sec486-SchwartzMS_BellmansSkor, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9873	0.9873	0.9873	0.9873
-UAS	0.7722	0.7722	0.7722	0.7722
-LAS	0.6582	0.6582	0.6582	0.6582
-
-Sample: svediakorp-sec330-GyllenborgC_SwenskaSpratthoken, time period: 1700-1750
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.8650	0.8650	0.8650	0.8650
-UAS	0.7178	0.7178	0.7178	0.7178
-LAS	0.6074	0.6074	0.6074	0.6074
-
-Sample: svediakorp-sec631-HasselskogN_HallaHallaGronkoping, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9605	0.9605	0.9605	0.9605
-UAS	0.8026	0.8026	0.8026	0.8026
-LAS	0.7105	0.7105	0.7105	0.7105
-
-Sample: svediakorp-sec277-EnbomPU_MedborgeligtSkalde, time period: 1750-1800
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9464	0.9464	0.9464	0.9464
-UAS	0.8214	0.8214	0.8214	0.8214
-LAS	0.7500	0.7500	0.7500	0.7500
-
-Sample: svediakorp-sec1102-spf259, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9444	0.9444	0.9444	0.9444
-UAS	0.8889	0.8889	0.8889	0.8889
-LAS	0.8889	0.8889	0.8889	0.8889
-
-Sample: svediakorp-sec991-spf148, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9561	0.9561	0.9561	0.9561
-UAS	0.7982	0.7982	0.7982	0.7982
-LAS	0.7544	0.7544	0.7544	0.7544
-
-Sample: svediakorp-sec639-HeidenstamV_Proletarfilosofiens, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9861	0.9861	0.9861	0.9861
-UAS	0.9306	0.9306	0.9306	0.9306
-LAS	0.8889	0.8889	0.8889	0.8889
-
-Sample: svediakorp-sec988-spf145, time period: 1800-1850
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9147	0.9147	0.9147	0.9147
-UAS	0.7442	0.7442	0.7442	0.7442
-LAS	0.6822	0.6822	0.6822	0.6822
-
-Sample: svediakorp-sec1063-spf220, time period: 1850-1900
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9660	0.9660	0.9660	0.9660
-UAS	0.8435	0.8435	0.8435	0.8435
-LAS	0.7551	0.7551	0.7551	0.7551
-
-Sample: svediakorp-sec1033-spf190, time period: 1900-1950
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9692	0.9692	0.9692	0.9692
-UAS	0.9077	0.9077	0.9077	0.9077
-LAS	0.8000	0.8000	0.8000	0.8000
-
-Sample: svediakorp-letter141673-Stalhammar, time period: 1700-1750
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.7059	0.7059	0.7059	0.7059
-UAS	0.6471	0.6471	0.6471	0.6471
-LAS	0.4706	0.4706	0.4706	0.4706
-
-----------------------------------------------------------------------
-Average scores for time period 1700-1750 (3 samples with a total of 8 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.8302	0.8302	0.8302	0.8302
-UAS	0.6693	0.6693	0.6693	0.6693
-LAS	0.5507	0.5507	0.5507	0.5507
-
-Average scores for time period 1750-1800 (2 samples with a total of 5 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9584	0.9584	0.9584	0.9584
-UAS	0.8218	0.8218	0.8218	0.8218
-LAS	0.7615	0.7615	0.7615	0.7615
-
-Average scores for time period 1800-1850 (6 samples with a total of 19 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9517	0.9517	0.9517	0.9517
-UAS	0.7615	0.7615	0.7615	0.7615
-LAS	0.6935	0.6935	0.6935	0.6935
-
-Average scores for time period 1850-1900 (5 samples with a total of 14 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9692	0.9692	0.9692	0.9692
-UAS	0.8210	0.8210	0.8210	0.8210
-LAS	0.7382	0.7382	0.7382	0.7382
-
-Average scores for time period 1900-1950 (5 samples with a total of 13 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9805	0.9805	0.9805	0.9805
-UAS	0.8734	0.8734	0.8734	0.8734
-LAS	0.8045	0.8045	0.8045	0.8045
-
-Overall scores across all time periods (21 samples with a total of 59 sentences):
-Metric	Precision	Recall	F1	Aligned Accuracy
-UPOS	0.9511	0.9511	0.9511	0.9511
-UAS	0.7878	0.7878	0.7878	0.7878
-LAS	0.7122	0.7122	0.7122	0.7122

logs/latest.txt

@@ -1 +0,0 @@
-log_bert-base-swedish-cased_sv_20260319_214621.txt

logs/log_bert-base-swedish-cased_sv_20260219_135555.txt

@@ -1,114 +0,0 @@
-=== LOGFILE: logs/log_bert-base-swedish-cased_sv_20260219_135555.txt ===
-Language codes: sv
-Using pretrained model: KBLab/bert-base-swedish-cased
-
-Running: python prepare-train-val-SILVER.py sv
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Loading DigPhil MACHINE for SILVER (minus gold overlap)…
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
-Adding 79 remaining gold sentences to TRAIN
-Cleaning TRAIN...
-Cleaning SILVER...
-[REMOVED] sent_id=11_12_13  ERRORS=['Multiple roots found: [5, 17, 25]']
-[REMOVED] sent_id=65_split2_66_split2  ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
-[REMOVED] sent_id=50_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
-[REMOVED] sent_id=53_54  ERRORS=['Multiple roots found: [27, 91]']
-[REMOVED] sent_id=55_56_57  ERRORS=['Multiple roots found: [2, 4, 13]']
-[REMOVED] sent_id=909  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=912  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=396  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=416  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=589  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=17_split1  ERRORS=['Multiple roots found: [2, 14, 17]']
-[REMOVED] sent_id=17_split2  ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
-[REMOVED] sent_id=19_split1  ERRORS=['Multiple roots found: [3, 31]']
-[REMOVED] sent_id=87_88  ERRORS=['Multiple roots found: [3, 6]']
-[REMOVED] sent_id=25  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=136  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=208  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=230  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=245  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=276  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=320  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=366  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=519  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=569  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=6  ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
-[REMOVED] sent_id=7_8  ERRORS=['Multiple roots found: [5, 10]']
-[REMOVED] sent_id=124_split2  ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
-[REMOVED] sent_id=30_31  ERRORS=['Multiple roots found: [3, 18]']
-[REMOVED] sent_id=35  ERRORS=['Line 36: Invalid token ID or head']
-[REMOVED] sent_id=119  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=188  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=223  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=268  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=325  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=388  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=399  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=475  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=505  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=520  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=562  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=711  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=731  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=867  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=884  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=923  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=939  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1086  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1251  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1345  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1459  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1656  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=3_split1  ERRORS=['Multiple roots found: [4, 15, 17]']
-[REMOVED] sent_id=3_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=3_4  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=5_6  ERRORS=['Multiple roots found: [3, 24]']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [7, 20]']
-[REMOVED] sent_id=8_9  ERRORS=['Multiple roots found: [24, 57]']
-[REMOVED] sent_id=12_13  ERRORS=['Multiple roots found: [11, 16]']
-Cleaning DEV...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Cleaning TEST...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Writing TRAIN → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (13873 valid sentences)
-Writing SILVER → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu (46432 valid sentences)
-Writing DEV → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (29 valid sentences)
-Writing TEST → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (all DIGPHIL_GOLD, 108 valid sentences)
-Done.
-Sourcing scripts/config.sh
-Running stanza dataset preparation…

logs/log_bert-base-swedish-cased_sv_20260219_135610.txt

@@ -1,138 +0,0 @@
-=== LOGFILE: logs/log_bert-base-swedish-cased_sv_20260219_135610.txt ===
-Language codes: sv
-Using pretrained model: KBLab/bert-base-swedish-cased
-
-Running: python prepare-train-val-SILVER.py sv
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Loading DigPhil MACHINE for SILVER (minus gold overlap)…
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
-Adding 75 remaining gold sentences to TRAIN
-Cleaning TRAIN...
-Cleaning SILVER...
-[REMOVED] sent_id=11_12_13  ERRORS=['Multiple roots found: [5, 17, 25]']
-[REMOVED] sent_id=65_split2_66_split2  ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
-[REMOVED] sent_id=50_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
-[REMOVED] sent_id=53_54  ERRORS=['Multiple roots found: [27, 91]']
-[REMOVED] sent_id=55_56_57  ERRORS=['Multiple roots found: [2, 4, 13]']
-[REMOVED] sent_id=909  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=912  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=396  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=416  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=589  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=17_split1  ERRORS=['Multiple roots found: [2, 14, 17]']
-[REMOVED] sent_id=17_split2  ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
-[REMOVED] sent_id=19_split1  ERRORS=['Multiple roots found: [3, 31]']
-[REMOVED] sent_id=87_88  ERRORS=['Multiple roots found: [3, 6]']
-[REMOVED] sent_id=25  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=136  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=208  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=230  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=245  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=276  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=320  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=366  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=519  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=569  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=6  ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
-[REMOVED] sent_id=7_8  ERRORS=['Multiple roots found: [5, 10]']
-[REMOVED] sent_id=124_split2  ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
-[REMOVED] sent_id=30_31  ERRORS=['Multiple roots found: [3, 18]']
-[REMOVED] sent_id=35  ERRORS=['Line 36: Invalid token ID or head']
-[REMOVED] sent_id=119  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=188  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=223  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=268  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=325  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=388  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=399  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=475  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=505  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=520  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=562  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=711  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=731  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=867  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=884  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=923  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=939  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1086  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1251  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1345  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1459  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1656  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=3_split1  ERRORS=['Multiple roots found: [4, 15, 17]']
-[REMOVED] sent_id=3_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=3_4  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=5_6  ERRORS=['Multiple roots found: [3, 24]']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [7, 20]']
-[REMOVED] sent_id=8_9  ERRORS=['Multiple roots found: [24, 57]']
-[REMOVED] sent_id=12_13  ERRORS=['Multiple roots found: [11, 16]']
-Cleaning DEV...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Cleaning TEST...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Writing TRAIN → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (13869 valid sentences)
-Writing SILVER → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu (46432 valid sentences)
-Writing DEV → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (33 valid sentences)
-Writing TEST → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (all DIGPHIL_GOLD, 108 valid sentences)
-Done.
-Sourcing scripts/config.sh
-Running stanza dataset preparation…
-2026-02-19 13:56:13 INFO: Datasets program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --gold
-Augmented 60 quotes: Counter({'《》': 9, '»«': 8, '″″': 8, '““': 7, '„”': 6, '""': 6, '„“': 6, '«»': 4, '””': 4, '「」': 2})
-Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /tmp/tmpzh8h6jmq/sv_diachronic.train.gold.conllu
-Swapped 'w1, w2' for 'w1 ,w2' 94 times
-Added 85 new sentences with asdf, zzzz -> asdf,zzzz
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /tmp/tmpzh8h6jmq/sv_diachronic.dev.gold.conllu
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /tmp/tmpzh8h6jmq/sv_diachronic.test.gold.conllu
-Copying from /tmp/tmpzh8h6jmq/sv_diachronic.train.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-Copying from /tmp/tmpzh8h6jmq/sv_diachronic.dev.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
-Copying from /tmp/tmpzh8h6jmq/sv_diachronic.test.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
-Running stanza dependency parser training…
-2026-02-19 13:56:16 INFO: Training program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --batch_size 32 --dropout 0.33 --use_bert --bert_model KBLab/bert-base-swedish-cased --silver_file ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu
-2026-02-19 13:56:16 DEBUG: UD_Swedish-diachronic: sv_diachronic
-2026-02-19 13:56:16 INFO: Using model /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2026-02-19 13:56:16 INFO: Using model /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2026-02-19 13:56:16 INFO: Using default pretrain for language sv, found in /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt  To use a different pretrain, specify --wordvec_pretrain_file
-2026-02-19 13:56:16 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_transformer_parser.pt exists, skipping!
-DONE.
-Full log saved to: logs/log_bert-base-swedish-cased_sv_20260219_135610.txt
-Symlink updated: logs/latest.txt → log_bert-base-swedish-cased_sv_20260219_135610.txt
-Latest log: log_bert-base-swedish-cased_sv_20260219_135610.txt

logs/log_bert-base-swedish-cased_sv_20260318_131623.txt

@@ -1,271 +0,0 @@
-=== LOGFILE: logs/log_bert-base-swedish-cased_sv_20260318_131623.txt ===
-Language codes: sv
-Using pretrained model: KBLab/bert-base-swedish-cased
-
-Running: python prepare-train-val-SILVER.py sv
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Loading DigPhil MACHINE for SILVER (minus gold overlap)…
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec991-spf148.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-letter141673-Stalhammar.conllu
-Adding 76 remaining gold sentences to TRAIN
-Cleaning TRAIN...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Cleaning SILVER...
-[REMOVED] sent_id=909  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=912  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=396  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=416  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=589  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=25  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=136  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=208  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=230  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=245  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=276  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=320  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=366  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=519  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=569  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=119  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=188  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=223  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=268  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=325  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=388  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=399  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=475  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=505  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=520  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=562  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=711  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=731  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=867  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=884  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=923  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=939  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1086  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1251  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1345  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1459  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1656  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=1266  ERRORS=['Token 17: Missing form']
-[REMOVED] sent_id=4897  ERRORS=['Token 46: Missing form', 'Token 47: Missing form']
-Cleaning DEV...
-Cleaning TEST...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Writing TRAIN → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (13869 valid sentences)
-Writing SILVER → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu (51431 valid sentences)
-Writing DEV → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (33 valid sentences)
-Writing TEST → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (all DIGPHIL_GOLD, 108 valid sentences)
-Done.
-Sourcing scripts/config.sh
-Running stanza dataset preparation…
-2026-03-18 13:16:26 INFO: Datasets program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --gold
-Augmented 60 quotes: Counter({'《》': 9, '»«': 8, '″″': 8, '““': 7, '„”': 6, '""': 6, '„“': 6, '«»': 4, '””': 4, '「」': 2})
-Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /tmp/tmpz3gm6g6g/sv_diachronic.train.gold.conllu
-Swapped 'w1, w2' for 'w1 ,w2' 94 times
-Added 85 new sentences with asdf, zzzz -> asdf,zzzz
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /tmp/tmpz3gm6g6g/sv_diachronic.dev.gold.conllu
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /tmp/tmpz3gm6g6g/sv_diachronic.test.gold.conllu
-Copying from /tmp/tmpz3gm6g6g/sv_diachronic.train.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-Copying from /tmp/tmpz3gm6g6g/sv_diachronic.dev.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
-Copying from /tmp/tmpz3gm6g6g/sv_diachronic.test.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
-Running stanza dependency parser training…
-2026-03-18 13:16:28 INFO: Training program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --batch_size 32 --dropout 0.33 --use_bert --bert_model KBLab/bert-base-swedish-cased --silver_file ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu
-2026-03-18 13:16:28 DEBUG: UD_Swedish-diachronic: sv_diachronic
-2026-03-18 13:16:28 INFO: Using model /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2026-03-18 13:16:28 INFO: Using model /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2026-03-18 13:16:28 INFO: Using default pretrain for language sv, found in /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt  To use a different pretrain, specify --wordvec_pretrain_file
-2026-03-18 13:16:28 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_transformer_parser.pt does not exist, training new model
-2026-03-18 13:16:28 INFO: Using model /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2026-03-18 13:16:28 INFO: Using model /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2026-03-18 13:16:28 INFO: Using default pretrain for language sv, found in /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt  To use a different pretrain, specify --wordvec_pretrain_file
-2026-03-18 13:16:28 INFO: Running train depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/home/urdatorn/stanza_resources/sv/pretrain', '--train_file', '/home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu', '--eval_file', '/home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--batch_size', '5000', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'train', '--wordvec_pretrain_file', '/home/urdatorn/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt', '--batch_size', '32', '--dropout', '0.33', '--bert_model', 'KBLab/bert-base-swedish-cased', '--silver_file', 'ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu']
-2026-03-18 13:16:28 INFO: Running parser in train mode
-2026-03-18 13:16:29 DEBUG: Loaded pretrain from /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt
-2026-03-18 13:16:29 INFO: Using 7 as the word cutoff based on the size of the pretrain (250000)
-2026-03-18 13:16:29 INFO: Using pretrained contextualized char embedding
-2026-03-18 13:16:29 INFO: ARGS USED AT TRAINING TIME:
-augment_nopunct: None
-batch_size: 32
-bert_finetune: False
-bert_finetune_layers: None
-bert_hidden_layers: 4
-bert_learning_rate: 1.0
-bert_model: KBLab/bert-base-swedish-cased
-bert_start_finetuning: 200
-bert_warmup_steps: 200
-bert_weight_decay: 0.0
-beta2: 0.999
-char: True
-char_emb_dim: 100
-char_hidden_dim: 400
-char_num_layers: 1
-char_rec_dropout: 0
-charlm: True
-charlm_backward_file: /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt
-charlm_forward_file: /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt
-charlm_save_dir: saved_models/charlm
-charlm_shorthand: sv_conll17
-checkpoint: True
-checkpoint_interval: 500
-checkpoint_save_name: None
-continue_from: None
-data_dir: data/depparse
-deep_biaff_hidden_dim: 400
-deep_biaff_output_dim: 160
-device: cpu
-distance: True
-dropout: 0.33
-eval_file: /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
-eval_interval: 100
-gold_labels: True
-hidden_dim: 400
-lang: sv
-linearization: True
-log_norms: False
-log_step: 20
-lora_alpha: 128
-lora_dropout: 0.1
-lora_modules_to_save: []
-lora_rank: 64
-lora_target_modules: ['query', 'value', 'output.dense', 'intermediate.dense']
-lr: 2.0
-max_grad_norm: 1.0
-max_steps: 50000
-max_steps_before_stop: 2000
-mode: train
-model_type: graph
-num_layers: 3
-optim: adadelta
-output_file: None
-output_latex: False
-pretrain: True
-pretrain_max_vocab: 250000
-rec_dropout: 0
-reversed: False
-sample_train: 1.0
-save_dir: saved_models/depparse
-save_name: {shorthand}_{embedding}_parser.pt
-second_batch_size: None
-second_bert_learning_rate: 0.001
-second_lr: 0.0002
-second_optim: adam
-second_optim_start_step: 10000
-second_warmup_steps: 200
-seed: 1234
-shorthand: sv_diachronic
-silver_file: ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu
-silver_weight: 0.5
-tag_emb_dim: 50
-train_file: /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-train_size: None
-transformed_dim: 125
-transition_embedding_dim: 20
-transition_hidden_dim: 20
-transition_merge_hidden_dim: 200
-transition_subtree_combination: SubtreeCombination.NONE
-transition_subtree_nonlinearity: none
-use_arc_embedding: False
-use_peft: False
-use_ufeats: True
-use_upos: True
-use_xpos: True
-wandb: False
-wandb_name: None
-weight_decay: 1e-05
-word_cutoff: 7
-word_dropout: 0.33
-word_emb_dim: 75
-wordvec_dir: /home/urdatorn/stanza_resources/sv/pretrain
-wordvec_file: None
-wordvec_pretrain_file: /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt
-
-2026-03-18 13:16:29 INFO: Loading data with batch size 32...
-2026-03-18 13:16:30 INFO: Train File /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu, Data Size: 13954
-2026-03-18 13:16:30 INFO: Original data size: 13954
-2026-03-18 13:16:34 WARNING: sv_diachronic is not a known dataset.  Examining the data to choose which xpos vocab to use
-2026-03-18 13:16:34 INFO: Original length = 13954
-2026-03-18 13:16:34 INFO: Filtered length = 13954
-2026-03-18 13:16:37 WARNING: Chose XPOSDescription(xpos_type=<XPOSType.XPOS: 1>, sep='-') for the xpos factory for sv_diachronic
-2026-03-18 13:16:41 DEBUG: 9513 batches created.
-2026-03-18 13:16:41 INFO: Augmented data size: 14712
-2026-03-18 13:16:47 DEBUG: 10020 batches created.
-2026-03-18 13:16:48 DEBUG: 32 batches created.
-2026-03-18 13:17:10 DEBUG: 28588 batches created.
-2026-03-18 13:17:10 INFO: Training parser...
-/home/urdatorn/.pyenv/versions/xpu/lib/python3.13/site-packages/huggingface_hub/file_download.py:949: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
-  warnings.warn(
-2026-03-18 13:17:12 DEBUG: Depparse model loading charmodels: /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt and /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt
-2026-03-18 13:17:12 DEBUG: Loading charlm from /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt
-2026-03-18 13:17:12 DEBUG: Loading charlm from /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt
-2026-03-18 13:17:13 DEBUG: Not using arc embedding enhancement
-2026-03-18 13:17:13 DEBUG: Building Adadelta with lr=2.000000, weight_decay=1e-05
-2026-03-18 13:17:29 INFO: Finished STEP 20/50000, loss = 2.292368 (0.203 sec/batch), lr: 2.000000
-2026-03-18 13:17:43 INFO: Finished STEP 40/50000, loss = 7.419842 (1.468 sec/batch), lr: 2.000000
-2026-03-18 13:17:51 INFO: Finished STEP 60/50000, loss = 7.826311 (1.111 sec/batch), lr: 2.000000
-2026-03-18 13:18:01 INFO: Finished STEP 80/50000, loss = 4.187514 (0.203 sec/batch), lr: 2.000000
-2026-03-18 13:18:09 INFO: Finished STEP 100/50000, loss = 3.687605 (0.196 sec/batch), lr: 2.000000
-2026-03-18 13:18:09 INFO: Evaluating on dev set...
-2026-03-18 13:18:16 INFO: LAS	MLAS	BLEX
-2026-03-18 13:18:16 INFO: 14.77	8.39	14.11
-2026-03-18 13:18:16 INFO: step 100: train_loss = 5.880744, dev_score = 0.1477
-2026-03-18 13:18:16 DEBUG:   training batch usage: Counter({0: 68, 1: 32})
-2026-03-18 13:18:16 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser.pt
-2026-03-18 13:18:16 INFO: new best model saved.
-2026-03-18 13:18:16 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser_checkpoint.pt
-2026-03-18 13:18:16 INFO: new model checkpoint saved.
-2026-03-18 13:18:27 INFO: Finished STEP 120/50000, loss = 5.895749 (0.951 sec/batch), lr: 2.000000
-2026-03-18 13:18:38 INFO: Finished STEP 140/50000, loss = 1.609854 (0.200 sec/batch), lr: 2.000000
-2026-03-18 13:18:45 INFO: Finished STEP 160/50000, loss = 1.833139 (0.220 sec/batch), lr: 2.000000
-2026-03-18 13:18:56 INFO: Finished STEP 180/50000, loss = 1.341930 (0.204 sec/batch), lr: 2.000000
-2026-03-18 13:19:04 INFO: Finished STEP 200/50000, loss = 2.078672 (0.224 sec/batch), lr: 2.000000
-2026-03-18 13:19:04 INFO: Evaluating on dev set...
-2026-03-18 13:19:11 INFO: LAS	MLAS	BLEX
-2026-03-18 13:19:11 INFO: 42.27	32.23	37.34
-2026-03-18 13:19:11 INFO: step 200: train_loss = 3.482911, dev_score = 0.4227
-2026-03-18 13:19:11 DEBUG:   training batch usage: Counter({0: 132, 1: 68})
-2026-03-18 13:19:11 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser.pt
-2026-03-18 13:19:11 INFO: new best model saved.
-2026-03-18 13:19:11 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser_checkpoint.pt
-2026-03-18 13:19:11 INFO: new model checkpoint saved.
-2026-03-18 13:19:19 INFO: Finished STEP 220/50000, loss = 1.613766 (0.210 sec/batch), lr: 2.000000
-2026-03-18 13:19:27 INFO: Finished STEP 240/50000, loss = 4.277971 (0.822 sec/batch), lr: 2.000000
-2026-03-18 13:19:39 INFO: Finished STEP 260/50000, loss = 4.801499 (0.823 sec/batch), lr: 2.000000
-2026-03-18 13:19:48 INFO: Finished STEP 280/50000, loss = 1.559014 (0.204 sec/batch), lr: 2.000000

logs/log_bert-base-swedish-cased_sv_20260318_132248.txt

@@ -1,109 +0,0 @@
-=== LOGFILE: logs/log_bert-base-swedish-cased_sv_20260318_132248.txt ===
-Language codes: sv
-Using pretrained model: KBLab/bert-base-swedish-cased
-
-Running: python prepare-train-val-SILVER.py sv
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Loading DigPhil MACHINE for SILVER (minus gold overlap)…
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec991-spf148.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-letter141673-Stalhammar.conllu
-Adding 77 remaining gold sentences to TRAIN
-Cleaning TRAIN...
-Cleaning SILVER...
-[REMOVED] sent_id=909  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=912  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=396  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=416  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=589  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=25  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=136  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=208  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=230  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=245  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=276  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=320  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=366  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=519  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=569  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=119  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=188  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=223  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=268  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=325  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=388  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=399  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=475  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=505  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=520  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=562  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=711  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=731  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=867  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=884  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=923  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=939  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1086  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1251  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1345  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1459  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1656  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=1266  ERRORS=['Token 17: Missing form']
-[REMOVED] sent_id=4897  ERRORS=['Token 46: Missing form', 'Token 47: Missing form']
-Cleaning DEV...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Cleaning TEST...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Writing TRAIN → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (13871 valid sentences)
-Writing SILVER → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu (51431 valid sentences)
-Writing DEV → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (31 valid sentences)
-Writing TEST → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (all DIGPHIL_GOLD, 108 valid sentences)
-Done.
-Sourcing scripts/config.sh
-Running stanza dataset preparation…
-2026-03-18 13:22:51 INFO: Datasets program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --gold
-Augmented 60 quotes: Counter({'《》': 9, '»«': 8, '″″': 8, '““': 7, '„”': 6, '""': 6, '„“': 6, '«»': 4, '””': 4, '「」': 2})
-Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /tmp/tmpw00sc1dk/sv_diachronic.train.gold.conllu
-Swapped 'w1, w2' for 'w1 ,w2' 94 times
-Added 86 new sentences with asdf, zzzz -> asdf,zzzz
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /tmp/tmpw00sc1dk/sv_diachronic.dev.gold.conllu
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /tmp/tmpw00sc1dk/sv_diachronic.test.gold.conllu
-Copying from /tmp/tmpw00sc1dk/sv_diachronic.train.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-Copying from /tmp/tmpw00sc1dk/sv_diachronic.dev.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
-Copying from /tmp/tmpw00sc1dk/sv_diachronic.test.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
-Running stanza dependency parser training…
-./make_new_model_BERT_SILVER.sh: line 72: unexpected EOF while looking for matching `"'

logs/log_bert-base-swedish-cased_sv_20260318_132404.txt

@@ -1,109 +0,0 @@
-=== LOGFILE: logs/log_bert-base-swedish-cased_sv_20260318_132404.txt ===
-Language codes: sv
-Using pretrained model: KBLab/bert-base-swedish-cased
-
-Running: python prepare-train-val-SILVER.py sv
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Loading DigPhil MACHINE for SILVER (minus gold overlap)…
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec991-spf148.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-letter141673-Stalhammar.conllu
-Adding 74 remaining gold sentences to TRAIN
-Cleaning TRAIN...
-Cleaning SILVER...
-[REMOVED] sent_id=909  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=912  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=396  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=416  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=589  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=25  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=136  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=208  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=230  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=245  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=276  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=320  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=366  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=519  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=569  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=119  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=188  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=223  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=268  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=325  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=388  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=399  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=475  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=505  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=520  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=562  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=711  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=731  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=867  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=884  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=923  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=939  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1086  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1251  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1345  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1459  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1656  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=1266  ERRORS=['Token 17: Missing form']
-[REMOVED] sent_id=4897  ERRORS=['Token 46: Missing form', 'Token 47: Missing form']
-Cleaning DEV...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Cleaning TEST...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Writing TRAIN → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (13868 valid sentences)
-Writing SILVER → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu (51431 valid sentences)
-Writing DEV → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (34 valid sentences)
-Writing TEST → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (all DIGPHIL_GOLD, 108 valid sentences)
-Done.
-Sourcing scripts/config.sh
-Running stanza dataset preparation…
-2026-03-18 13:24:07 INFO: Datasets program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --gold
-Augmented 60 quotes: Counter({'《》': 9, '»«': 8, '″″': 8, '““': 7, '„”': 6, '""': 6, '„“': 6, '«»': 4, '””': 4, '「」': 2})
-Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /tmp/tmp0l99psu3/sv_diachronic.train.gold.conllu
-Swapped 'w1, w2' for 'w1 ,w2' 93 times
-Added 85 new sentences with asdf, zzzz -> asdf,zzzz
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /tmp/tmp0l99psu3/sv_diachronic.dev.gold.conllu
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /tmp/tmp0l99psu3/sv_diachronic.test.gold.conllu
-Copying from /tmp/tmp0l99psu3/sv_diachronic.train.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-Copying from /tmp/tmp0l99psu3/sv_diachronic.dev.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
-Copying from /tmp/tmp0l99psu3/sv_diachronic.test.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
-Running stanza dependency parser training…
-./make_new_model_BERT_SILVER.sh: line 72: unexpected EOF while looking for matching `"'

logs/log_bert-base-swedish-cased_sv_20260318_132733.txt

@@ -1,325 +0,0 @@
-=== LOGFILE: logs/log_bert-base-swedish-cased_sv_20260318_132733.txt ===
-Language codes: sv
-Using pretrained model: KBLab/bert-base-swedish-cased
-
-Running: python prepare-train-val-SILVER.py sv
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Loading DigPhil MACHINE for SILVER (minus gold overlap)…
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec991-spf148.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /home/urdatorn/git/Diachronic_Treebanks_DigPhil/parsed_data/validated/svediakorp-letter141673-Stalhammar.conllu
-Adding 73 remaining gold sentences to TRAIN
-Cleaning TRAIN...
-Cleaning SILVER...
-[REMOVED] sent_id=909  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=912  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=396  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=416  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=589  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=25  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=136  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=208  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=230  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=245  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=276  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=320  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=366  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=519  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=569  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=119  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=188  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=223  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=268  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=325  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=388  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=399  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=475  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=505  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=520  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=562  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=711  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=731  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=867  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=884  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=923  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=939  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1086  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1251  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1345  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1459  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1656  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=1266  ERRORS=['Token 17: Missing form']
-[REMOVED] sent_id=4897  ERRORS=['Token 46: Missing form', 'Token 47: Missing form']
-Cleaning DEV...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Cleaning TEST...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Writing TRAIN → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (13867 valid sentences)
-Writing SILVER → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu (51431 valid sentences)
-Writing DEV → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (35 valid sentences)
-Writing TEST → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (all DIGPHIL_GOLD, 108 valid sentences)
-Done.
-Sourcing scripts/config.sh
-Running stanza dataset preparation…
-2026-03-18 13:27:36 INFO: Datasets program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --gold
-Augmented 60 quotes: Counter({'《》': 9, '»«': 8, '″″': 8, '““': 7, '„”': 6, '""': 6, '„“': 6, '«»': 4, '””': 4, '「」': 2})
-Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /tmp/tmpl2s5qxw1/sv_diachronic.train.gold.conllu
-Swapped 'w1, w2' for 'w1 ,w2' 98 times
-Added 85 new sentences with asdf, zzzz -> asdf,zzzz
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /tmp/tmpl2s5qxw1/sv_diachronic.dev.gold.conllu
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /tmp/tmpl2s5qxw1/sv_diachronic.test.gold.conllu
-Copying from /tmp/tmpl2s5qxw1/sv_diachronic.train.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-Copying from /tmp/tmpl2s5qxw1/sv_diachronic.dev.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
-Copying from /tmp/tmpl2s5qxw1/sv_diachronic.test.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
-Running stanza dependency parser training…
-2026-03-18 13:27:38 INFO: Training program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --batch_size 32 --dropout 0.33 --use_bert --bert_model KBLab/bert-base-swedish-cased --silver_file ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu --device xpu:0
-2026-03-18 13:27:38 DEBUG: UD_Swedish-diachronic: sv_diachronic
-2026-03-18 13:27:38 INFO: Using model /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2026-03-18 13:27:38 INFO: Using model /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2026-03-18 13:27:38 INFO: Using default pretrain for language sv, found in /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt  To use a different pretrain, specify --wordvec_pretrain_file
-2026-03-18 13:27:38 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_transformer_parser.pt does not exist, training new model
-2026-03-18 13:27:38 INFO: Using model /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2026-03-18 13:27:38 INFO: Using model /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2026-03-18 13:27:38 INFO: Using default pretrain for language sv, found in /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt  To use a different pretrain, specify --wordvec_pretrain_file
-2026-03-18 13:27:38 INFO: Running train depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/home/urdatorn/stanza_resources/sv/pretrain', '--train_file', '/home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu', '--eval_file', '/home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--batch_size', '5000', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'train', '--wordvec_pretrain_file', '/home/urdatorn/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt', '--batch_size', '32', '--dropout', '0.33', '--bert_model', 'KBLab/bert-base-swedish-cased', '--silver_file', 'ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu', '--device', 'xpu:0']
-2026-03-18 13:27:38 INFO: Running parser in train mode
-2026-03-18 13:27:39 DEBUG: Loaded pretrain from /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt
-2026-03-18 13:27:39 INFO: Using 7 as the word cutoff based on the size of the pretrain (250000)
-2026-03-18 13:27:39 INFO: Using pretrained contextualized char embedding
-2026-03-18 13:27:39 INFO: ARGS USED AT TRAINING TIME:
-augment_nopunct: None
-batch_size: 32
-bert_finetune: False
-bert_finetune_layers: None
-bert_hidden_layers: 4
-bert_learning_rate: 1.0
-bert_model: KBLab/bert-base-swedish-cased
-bert_start_finetuning: 200
-bert_warmup_steps: 200
-bert_weight_decay: 0.0
-beta2: 0.999
-char: True
-char_emb_dim: 100
-char_hidden_dim: 400
-char_num_layers: 1
-char_rec_dropout: 0
-charlm: True
-charlm_backward_file: /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt
-charlm_forward_file: /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt
-charlm_save_dir: saved_models/charlm
-charlm_shorthand: sv_conll17
-checkpoint: True
-checkpoint_interval: 500
-checkpoint_save_name: None
-continue_from: None
-data_dir: data/depparse
-deep_biaff_hidden_dim: 400
-deep_biaff_output_dim: 160
-device: xpu:0
-distance: True
-dropout: 0.33
-eval_file: /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
-eval_interval: 100
-gold_labels: True
-hidden_dim: 400
-lang: sv
-linearization: True
-log_norms: False
-log_step: 20
-lora_alpha: 128
-lora_dropout: 0.1
-lora_modules_to_save: []
-lora_rank: 64
-lora_target_modules: ['query', 'value', 'output.dense', 'intermediate.dense']
-lr: 2.0
-max_grad_norm: 1.0
-max_steps: 50000
-max_steps_before_stop: 2000
-mode: train
-model_type: graph
-num_layers: 3
-optim: adadelta
-output_file: None
-output_latex: False
-pretrain: True
-pretrain_max_vocab: 250000
-rec_dropout: 0
-reversed: False
-sample_train: 1.0
-save_dir: saved_models/depparse
-save_name: {shorthand}_{embedding}_parser.pt
-second_batch_size: None
-second_bert_learning_rate: 0.001
-second_lr: 0.0002
-second_optim: adam
-second_optim_start_step: 10000
-second_warmup_steps: 200
-seed: 1234
-shorthand: sv_diachronic
-silver_file: ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu
-silver_weight: 0.5
-tag_emb_dim: 50
-train_file: /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-train_size: None
-transformed_dim: 125
-transition_embedding_dim: 20
-transition_hidden_dim: 20
-transition_merge_hidden_dim: 200
-transition_subtree_combination: SubtreeCombination.NONE
-transition_subtree_nonlinearity: none
-use_arc_embedding: False
-use_peft: False
-use_ufeats: True
-use_upos: True
-use_xpos: True
-wandb: False
-wandb_name: None
-weight_decay: 1e-05
-word_cutoff: 7
-word_dropout: 0.33
-word_emb_dim: 75
-wordvec_dir: /home/urdatorn/stanza_resources/sv/pretrain
-wordvec_file: None
-wordvec_pretrain_file: /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt
-
-2026-03-18 13:27:39 INFO: Loading data with batch size 32...
-2026-03-18 13:27:40 INFO: Train File /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu, Data Size: 13952
-2026-03-18 13:27:40 INFO: Original data size: 13952
-2026-03-18 13:27:43 WARNING: sv_diachronic is not a known dataset.  Examining the data to choose which xpos vocab to use
-2026-03-18 13:27:43 INFO: Original length = 13952
-2026-03-18 13:27:43 INFO: Filtered length = 13952
-2026-03-18 13:27:47 WARNING: Chose XPOSDescription(xpos_type=<XPOSType.XPOS: 1>, sep='-') for the xpos factory for sv_diachronic
-2026-03-18 13:27:50 DEBUG: 9512 batches created.
-2026-03-18 13:27:50 INFO: Augmented data size: 14707
-2026-03-18 13:27:57 DEBUG: 10016 batches created.
-2026-03-18 13:27:57 DEBUG: 34 batches created.
-2026-03-18 13:28:19 DEBUG: 28588 batches created.
-2026-03-18 13:28:19 INFO: Training parser...
-/home/urdatorn/.pyenv/versions/xpu/lib/python3.13/site-packages/huggingface_hub/file_download.py:949: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
-  warnings.warn(
-2026-03-18 13:28:20 DEBUG: Depparse model loading charmodels: /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt and /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt
-2026-03-18 13:28:20 DEBUG: Loading charlm from /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt
-2026-03-18 13:28:20 DEBUG: Loading charlm from /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt
-2026-03-18 13:28:20 DEBUG: Not using arc embedding enhancement
-2026-03-18 13:28:21 DEBUG: Building Adadelta with lr=2.000000, weight_decay=1e-05
-2026-03-18 13:28:31 INFO: Finished STEP 20/50000, loss = 5.042704 (0.060 sec/batch), lr: 2.000000
-2026-03-18 13:28:38 INFO: Finished STEP 40/50000, loss = 7.420871 (0.629 sec/batch), lr: 2.000000
-2026-03-18 13:28:42 INFO: Finished STEP 60/50000, loss = 7.130751 (0.585 sec/batch), lr: 2.000000
-2026-03-18 13:28:46 INFO: Finished STEP 80/50000, loss = 2.640507 (0.061 sec/batch), lr: 2.000000
-2026-03-18 13:28:49 INFO: Finished STEP 100/50000, loss = 2.198331 (0.062 sec/batch), lr: 2.000000
-2026-03-18 13:28:49 INFO: Evaluating on dev set...
-2026-03-18 13:28:52 INFO: LAS	MLAS	BLEX
-2026-03-18 13:28:52 INFO: 14.92	7.39	9.17
-2026-03-18 13:28:52 INFO: step 100: train_loss = 6.003559, dev_score = 0.1492
-2026-03-18 13:28:52 DEBUG:   training batch usage: Counter({0: 68, 1: 32})
-2026-03-18 13:28:52 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser.pt
-2026-03-18 13:28:52 INFO: new best model saved.
-2026-03-18 13:28:52 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser_checkpoint.pt
-2026-03-18 13:28:52 INFO: new model checkpoint saved.
-2026-03-18 13:28:58 INFO: Finished STEP 120/50000, loss = 5.791277 (0.479 sec/batch), lr: 2.000000
-2026-03-18 13:29:02 INFO: Finished STEP 140/50000, loss = 1.561704 (0.065 sec/batch), lr: 2.000000
-2026-03-18 13:29:05 INFO: Finished STEP 160/50000, loss = 2.749855 (0.064 sec/batch), lr: 2.000000
-2026-03-18 13:29:10 INFO: Finished STEP 180/50000, loss = 2.644507 (0.062 sec/batch), lr: 2.000000
-2026-03-18 13:29:14 INFO: Finished STEP 200/50000, loss = 2.221299 (0.062 sec/batch), lr: 2.000000
-2026-03-18 13:29:14 INFO: Evaluating on dev set...
-2026-03-18 13:29:17 INFO: LAS	MLAS	BLEX
-2026-03-18 13:29:17 INFO: 36.67	23.23	32.82
-2026-03-18 13:29:17 INFO: step 200: train_loss = 3.408124, dev_score = 0.3667
-2026-03-18 13:29:17 DEBUG:   training batch usage: Counter({0: 132, 1: 68})
-2026-03-18 13:29:17 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser.pt
-2026-03-18 13:29:17 INFO: new best model saved.
-2026-03-18 13:29:17 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser_checkpoint.pt
-2026-03-18 13:29:17 INFO: new model checkpoint saved.
-2026-03-18 13:29:20 INFO: Finished STEP 220/50000, loss = 2.403125 (0.064 sec/batch), lr: 2.000000
-2026-03-18 13:29:24 INFO: Finished STEP 240/50000, loss = 3.965761 (0.412 sec/batch), lr: 2.000000
-2026-03-18 13:29:30 INFO: Finished STEP 260/50000, loss = 4.813926 (0.403 sec/batch), lr: 2.000000
-2026-03-18 13:29:33 INFO: Finished STEP 280/50000, loss = 2.433779 (0.067 sec/batch), lr: 2.000000
-2026-03-18 13:29:37 INFO: Finished STEP 300/50000, loss = 5.255473 (0.374 sec/batch), lr: 2.000000
-2026-03-18 13:29:37 INFO: Evaluating on dev set...
-2026-03-18 13:29:40 INFO: LAS	MLAS	BLEX
-2026-03-18 13:29:40 INFO: 51.20	39.90	43.72
-2026-03-18 13:29:40 INFO: step 300: train_loss = 3.172424, dev_score = 0.5120
-2026-03-18 13:29:40 DEBUG:   training batch usage: Counter({0: 191, 1: 109})
-2026-03-18 13:29:40 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser.pt
-2026-03-18 13:29:40 INFO: new best model saved.
-2026-03-18 13:29:41 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser_checkpoint.pt
-2026-03-18 13:29:41 INFO: new model checkpoint saved.
-2026-03-18 13:29:42 INFO: Finished STEP 320/50000, loss = 4.718216 (0.382 sec/batch), lr: 2.000000
-2026-03-18 13:29:46 INFO: Finished STEP 340/50000, loss = 2.242086 (0.065 sec/batch), lr: 2.000000
-2026-03-18 13:29:48 INFO: Finished STEP 360/50000, loss = 1.605185 (0.065 sec/batch), lr: 2.000000
-2026-03-18 13:29:50 INFO: Finished STEP 380/50000, loss = 1.821524 (0.063 sec/batch), lr: 2.000000
-2026-03-18 13:29:54 INFO: Finished STEP 400/50000, loss = 1.597555 (0.065 sec/batch), lr: 2.000000
-2026-03-18 13:29:54 INFO: Evaluating on dev set...
-2026-03-18 13:29:57 INFO: LAS	MLAS	BLEX
-2026-03-18 13:29:57 INFO: 50.46	38.59	43.93
-2026-03-18 13:29:57 INFO: step 400: train_loss = 2.549639, dev_score = 0.5046
-2026-03-18 13:29:57 DEBUG:   training batch usage: Counter({0: 269, 1: 131})
-2026-03-18 13:29:57 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser_checkpoint.pt
-2026-03-18 13:29:57 INFO: new model checkpoint saved.
-2026-03-18 13:30:00 INFO: Finished STEP 420/50000, loss = 2.220649 (0.065 sec/batch), lr: 2.000000
-2026-03-18 13:30:04 INFO: Finished STEP 440/50000, loss = 1.550643 (0.064 sec/batch), lr: 2.000000
-2026-03-18 13:30:08 INFO: Finished STEP 460/50000, loss = 3.319951 (0.344 sec/batch), lr: 2.000000
-2026-03-18 13:30:12 INFO: Finished STEP 480/50000, loss = 2.615420 (0.068 sec/batch), lr: 2.000000
-2026-03-18 13:30:14 INFO: Finished STEP 500/50000, loss = 1.976453 (0.068 sec/batch), lr: 2.000000
-2026-03-18 13:30:14 INFO: Evaluating on dev set...
-2026-03-18 13:30:17 INFO: LAS	MLAS	BLEX
-2026-03-18 13:30:17 INFO: 52.11	42.78	46.03
-2026-03-18 13:30:17 INFO: step 500: train_loss = 2.721934, dev_score = 0.5211
-2026-03-18 13:30:17 DEBUG:   training batch usage: Counter({0: 332, 1: 168})
-2026-03-18 13:30:17 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser.pt
-2026-03-18 13:30:17 INFO: new best model saved.
-2026-03-18 13:30:18 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser_checkpoint.pt
-2026-03-18 13:30:18 INFO: new model checkpoint saved.
-2026-03-18 13:30:21 INFO: Finished STEP 520/50000, loss = 4.217991 (0.338 sec/batch), lr: 2.000000
-2026-03-18 13:30:24 INFO: Finished STEP 540/50000, loss = 3.852498 (0.337 sec/batch), lr: 2.000000
-2026-03-18 13:30:28 INFO: Finished STEP 560/50000, loss = 1.689470 (0.069 sec/batch), lr: 2.000000
-2026-03-18 13:30:32 INFO: Finished STEP 580/50000, loss = 3.136248 (0.339 sec/batch), lr: 2.000000
-2026-03-18 13:30:35 INFO: Finished STEP 600/50000, loss = 3.013068 (0.068 sec/batch), lr: 2.000000
-2026-03-18 13:30:35 INFO: Evaluating on dev set...
-2026-03-18 13:30:38 INFO: LAS	MLAS	BLEX
-2026-03-18 13:30:38 INFO: 56.89	46.97	50.02
-2026-03-18 13:30:38 INFO: step 600: train_loss = 2.802572, dev_score = 0.5689
-2026-03-18 13:30:38 DEBUG:   training batch usage: Counter({0: 392, 1: 208})
-2026-03-18 13:30:38 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser.pt
-2026-03-18 13:30:38 INFO: new best model saved.
-2026-03-18 13:30:38 INFO: Model saved to saved_models/depparse/sv_diachronic_transformer_parser_checkpoint.pt
-2026-03-18 13:30:38 INFO: new model checkpoint saved.
-2026-03-18 13:30:41 INFO: Finished STEP 620/50000, loss = 3.173692 (0.327 sec/batch), lr: 2.000000
-2026-03-18 13:30:44 INFO: Finished STEP 640/50000, loss = 1.843263 (0.070 sec/batch), lr: 2.000000
-2026-03-18 13:30:47 INFO: Finished STEP 660/50000, loss = 1.417565 (0.069 sec/batch), lr: 2.000000
-2026-03-18 13:30:51 INFO: Finished STEP 680/50000, loss = 4.149919 (0.325 sec/batch), lr: 2.000000

logs/log_conll17.pt_sv_20251203_232257.txt

@@ -1,719 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_20251203_232257.txt ===
-Language codes: sv
-Using pretrained model: conll17.pt
-
-Running: python prepare-train-val-test.py sv
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/ucxn_ud_swedish-talbanken.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Skipping DigPhil MACHINE (diachron not requested).
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
-Cleaning TRAIN...
-Cleaning DEV...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Cleaning TEST...
-Writing TRAIN → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (19820 valid sentences)
-Writing DEV → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (9 valid sentences)
-Writing TEST → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (99 valid sentences)
-Done.
-Sourcing scripts/config_alvis.sh
-Running stanza dataset preparation…
-2025-12-03 23:23:04 INFO: Datasets program called with:
-/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 23:23:04 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
-
-2025-12-03 23:23:04 INFO: Downloaded file to /cephyr/users/cleland/Alvis/stanza_resources/resources.json
-2025-12-03 23:23:04 DEBUG: Processing parameter "processors"...
-2025-12-03 23:23:04 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
-2025-12-03 23:23:04 INFO: Downloading these customized packages for language: sv (Swedish)...
-=======================
-| Processor | Package |
------------------------
-=======================
-
-2025-12-03 23:23:04 INFO: Finished downloading models and saved to /cephyr/users/cleland/Alvis/stanza_resources
-2025-12-03 23:23:04 INFO: Using tagger model in /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt for sv_diachronic
-2025-12-03 23:23:04 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2025-12-03 23:23:04 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-Augmented 56 quotes: Counter({'„”': 9, '″″': 9, '""': 8, '「」': 8, '””': 5, '““': 4, '《》': 4, '»«': 3, '„“': 3, '«»': 3})
-2025-12-03 23:23:05 INFO: Running tagger to retag /local/tmp.5441282/tmp4sg9id1k/sv_diachronic.train.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5441282/tmp4sg9id1k/sv_diachronic.train.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu']
-2025-12-03 23:23:05 INFO: Running tagger in predict mode
-2025-12-03 23:23:05 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2025-12-03 23:23:07 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 23:23:07 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 23:23:07 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 23:23:07 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 23:23:07 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 23:23:10 INFO: Loading data with batch size 250...
-2025-12-03 23:23:24 INFO: Start evaluation...
-2025-12-03 23:24:31 INFO: UPOS	XPOS	UFeats	AllTags
-2025-12-03 23:24:31 INFO: 98.01	62.93	94.04	60.49
-2025-12-03 23:24:31 INFO: POS Tagger score: sv_diachronic 60.49
-2025-12-03 23:24:32 INFO: Running tagger to retag /local/tmp.5441282/tmp4sg9id1k/sv_diachronic.dev.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5441282/tmp4sg9id1k/sv_diachronic.dev.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu']
-2025-12-03 23:24:32 INFO: Running tagger in predict mode
-2025-12-03 23:24:32 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2025-12-03 23:24:33 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 23:24:33 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 23:24:33 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 23:24:33 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 23:24:34 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 23:24:34 INFO: Loading data with batch size 250...
-2025-12-03 23:24:34 INFO: Start evaluation...
-2025-12-03 23:24:34 INFO: UPOS	XPOS	UFeats	AllTags
-2025-12-03 23:24:34 INFO: 93.32	90.84	93.32	85.64
-2025-12-03 23:24:34 INFO: POS Tagger score: sv_diachronic 85.64
-2025-12-03 23:24:34 INFO: Running tagger to retag /local/tmp.5441282/tmp4sg9id1k/sv_diachronic.test.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5441282/tmp4sg9id1k/sv_diachronic.test.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu']
-2025-12-03 23:24:34 INFO: Running tagger in predict mode
-2025-12-03 23:24:34 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2025-12-03 23:24:36 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 23:24:36 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 23:24:36 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 23:24:36 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 23:24:36 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 23:24:36 INFO: Loading data with batch size 250...
-2025-12-03 23:24:36 INFO: Start evaluation...
-2025-12-03 23:24:37 INFO: UPOS	XPOS	UFeats	AllTags
-2025-12-03 23:24:37 INFO: 93.14	96.78	95.32	90.28
-2025-12-03 23:24:37 INFO: POS Tagger score: sv_diachronic 90.28
-Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /local/tmp.5441282/tmp4sg9id1k/sv_diachronic.train.gold.conllu
-Swapped 'w1, w2' for 'w1 ,w2' 122 times
-Added 100 new sentences with asdf, zzzz -> asdf,zzzz
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /local/tmp.5441282/tmp4sg9id1k/sv_diachronic.dev.gold.conllu
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /local/tmp.5441282/tmp4sg9id1k/sv_diachronic.test.gold.conllu
-Running stanza dependency parser training…
-2025-12-03 23:24:46 INFO: Training program called with:
-/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt --batch_size 32 --dropout 0.33
-2025-12-03 23:24:46 DEBUG: UD_Swedish-diachronic: sv_diachronic
-2025-12-03 23:24:46 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2025-12-03 23:24:46 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2025-12-03 23:24:46 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_charlm_parser.pt does not exist, training new model
-2025-12-03 23:24:46 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2025-12-03 23:24:46 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2025-12-03 23:24:46 INFO: Running train depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--train_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--batch_size', '5000', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'train', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
-2025-12-03 23:24:46 INFO: Running parser in train mode
-2025-12-03 23:24:46 INFO: Using pretrained contextualized char embedding
-2025-12-03 23:24:46 INFO: Loading data with batch size 32...
-2025-12-03 23:24:49 INFO: Train File /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu, Data Size: 19920
-2025-12-03 23:24:49 INFO: Original data size: 19920
-2025-12-03 23:24:49 INFO: Augmented data size: 20844
-2025-12-03 23:24:57 WARNING: sv_diachronic is not a known dataset.  Examining the data to choose which xpos vocab to use
-2025-12-03 23:24:57 INFO: Original length = 20844
-2025-12-03 23:24:57 INFO: Filtered length = 20844
-2025-12-03 23:25:02 WARNING: Chose XPOSDescription(xpos_type=<XPOSType.XPOS: 1>, sep='|') for the xpos factory for sv_diachronic
-2025-12-03 23:25:06 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 23:25:11 DEBUG: 13986 batches created.
-2025-12-03 23:25:11 DEBUG: 9 batches created.
-2025-12-03 23:25:11 INFO: Training parser...
-2025-12-03 23:25:11 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 23:25:11 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 23:25:11 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 23:25:11 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 23:25:14 INFO: Finished STEP 20/50000, loss = 3.699069 (0.033 sec/batch), lr: 0.003000
-2025-12-03 23:25:15 INFO: Finished STEP 40/50000, loss = 3.841336 (0.035 sec/batch), lr: 0.003000
-2025-12-03 23:25:16 INFO: Finished STEP 60/50000, loss = 2.719965 (0.034 sec/batch), lr: 0.003000
-2025-12-03 23:25:16 INFO: Finished STEP 80/50000, loss = 3.832992 (0.035 sec/batch), lr: 0.003000
-2025-12-03 23:25:17 INFO: Finished STEP 100/50000, loss = 3.050168 (0.035 sec/batch), lr: 0.003000
-2025-12-03 23:25:17 INFO: Evaluating on dev set...
-2025-12-03 23:25:18 INFO: LAS	MLAS	BLEX
-2025-12-03 23:25:18 INFO: 3.22	0.82	1.64
-2025-12-03 23:25:18 INFO: step 100: train_loss = 1815.927894, dev_score = 0.0322
-2025-12-03 23:25:18 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:25:18 INFO: new best model saved.
-2025-12-03 23:25:19 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:25:19 INFO: new model checkpoint saved.
-2025-12-03 23:25:19 INFO: Finished STEP 120/50000, loss = 3.082294 (0.035 sec/batch), lr: 0.003000
-2025-12-03 23:25:20 INFO: Finished STEP 140/50000, loss = 2.829382 (0.035 sec/batch), lr: 0.003000
-2025-12-03 23:25:21 INFO: Finished STEP 160/50000, loss = 2.907810 (0.036 sec/batch), lr: 0.003000
-2025-12-03 23:25:22 INFO: Finished STEP 180/50000, loss = 3.411438 (0.035 sec/batch), lr: 0.003000
-2025-12-03 23:25:22 INFO: Finished STEP 200/50000, loss = 2.634063 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:25:22 INFO: Evaluating on dev set...
-2025-12-03 23:25:23 INFO: LAS	MLAS	BLEX
-2025-12-03 23:25:23 INFO: 7.92	6.14	7.51
-2025-12-03 23:25:23 INFO: step 200: train_loss = 2.962158, dev_score = 0.0792
-2025-12-03 23:25:23 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:25:23 INFO: new best model saved.
-2025-12-03 23:25:24 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:25:24 INFO: new model checkpoint saved.
-2025-12-03 23:25:25 INFO: Finished STEP 220/50000, loss = 2.243211 (0.036 sec/batch), lr: 0.003000
-2025-12-03 23:25:25 INFO: Finished STEP 240/50000, loss = 3.037997 (0.034 sec/batch), lr: 0.003000
-2025-12-03 23:25:26 INFO: Finished STEP 260/50000, loss = 2.535078 (0.034 sec/batch), lr: 0.003000
-2025-12-03 23:25:27 INFO: Finished STEP 280/50000, loss = 3.026233 (0.035 sec/batch), lr: 0.003000
-2025-12-03 23:25:27 INFO: Finished STEP 300/50000, loss = 3.630493 (0.034 sec/batch), lr: 0.003000
-2025-12-03 23:25:27 INFO: Evaluating on dev set...
-2025-12-03 23:25:28 INFO: LAS	MLAS	BLEX
-2025-12-03 23:25:28 INFO: 13.61	10.54	12.24
-2025-12-03 23:25:28 INFO: step 300: train_loss = 3.006871, dev_score = 0.1361
-2025-12-03 23:25:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:25:28 INFO: new best model saved.
-2025-12-03 23:25:29 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:25:29 INFO: new model checkpoint saved.
-2025-12-03 23:25:30 INFO: Finished STEP 320/50000, loss = 2.784619 (0.034 sec/batch), lr: 0.003000
-2025-12-03 23:25:30 INFO: Finished STEP 340/50000, loss = 3.004477 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:25:31 INFO: Finished STEP 360/50000, loss = 2.542431 (0.035 sec/batch), lr: 0.003000
-2025-12-03 23:25:32 INFO: Finished STEP 380/50000, loss = 3.084781 (0.036 sec/batch), lr: 0.003000
-2025-12-03 23:25:32 INFO: Finished STEP 400/50000, loss = 2.454229 (0.035 sec/batch), lr: 0.003000
-2025-12-03 23:25:32 INFO: Evaluating on dev set...
-2025-12-03 23:25:33 INFO: LAS	MLAS	BLEX
-2025-12-03 23:25:33 INFO: 18.81	6.64	10.55
-2025-12-03 23:25:33 INFO: step 400: train_loss = 3.048551, dev_score = 0.1881
-2025-12-03 23:25:33 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:25:33 INFO: new best model saved.
-2025-12-03 23:25:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:25:34 INFO: new model checkpoint saved.
-2025-12-03 23:25:35 INFO: Finished STEP 420/50000, loss = 2.214305 (0.034 sec/batch), lr: 0.003000
-2025-12-03 23:25:35 INFO: Finished STEP 440/50000, loss = 2.953549 (0.034 sec/batch), lr: 0.003000
-2025-12-03 23:25:36 INFO: Finished STEP 460/50000, loss = 2.711811 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:25:37 INFO: Finished STEP 480/50000, loss = 2.823795 (0.036 sec/batch), lr: 0.003000
-2025-12-03 23:25:38 INFO: Finished STEP 500/50000, loss = 4.227708 (0.035 sec/batch), lr: 0.003000
-2025-12-03 23:25:38 INFO: Evaluating on dev set...
-2025-12-03 23:25:38 INFO: LAS	MLAS	BLEX
-2025-12-03 23:25:38 INFO: 18.07	7.92	9.43
-2025-12-03 23:25:38 INFO: step 500: train_loss = 3.085661, dev_score = 0.1807
-2025-12-03 23:25:39 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:25:39 INFO: new model checkpoint saved.
-2025-12-03 23:25:39 INFO: Finished STEP 520/50000, loss = 3.152939 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:25:40 INFO: Finished STEP 540/50000, loss = 2.722816 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:25:41 INFO: Finished STEP 560/50000, loss = 1.797972 (0.034 sec/batch), lr: 0.003000
-2025-12-03 23:25:42 INFO: Finished STEP 580/50000, loss = 1.902476 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:25:42 INFO: Finished STEP 600/50000, loss = 2.532953 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:25:42 INFO: Evaluating on dev set...
-2025-12-03 23:25:43 INFO: LAS	MLAS	BLEX
-2025-12-03 23:25:43 INFO: 24.75	12.64	15.33
-2025-12-03 23:25:43 INFO: step 600: train_loss = 3.006814, dev_score = 0.2475
-2025-12-03 23:25:43 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:25:43 INFO: new best model saved.
-2025-12-03 23:25:44 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:25:44 INFO: new model checkpoint saved.
-2025-12-03 23:25:44 INFO: Finished STEP 620/50000, loss = 3.911072 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:25:45 INFO: Finished STEP 640/50000, loss = 3.704555 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:25:46 INFO: Finished STEP 660/50000, loss = 2.692690 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:25:47 INFO: Finished STEP 680/50000, loss = 2.771069 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:25:47 INFO: Finished STEP 700/50000, loss = 4.281591 (0.036 sec/batch), lr: 0.003000
-2025-12-03 23:25:47 INFO: Evaluating on dev set...
-2025-12-03 23:25:48 INFO: LAS	MLAS	BLEX
-2025-12-03 23:25:48 INFO: 34.65	21.99	25.31
-2025-12-03 23:25:48 INFO: step 700: train_loss = 3.216450, dev_score = 0.3465
-2025-12-03 23:25:48 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:25:48 INFO: new best model saved.
-2025-12-03 23:25:49 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:25:49 INFO: new model checkpoint saved.
-2025-12-03 23:25:50 INFO: Finished STEP 720/50000, loss = 3.350647 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:25:50 INFO: Finished STEP 740/50000, loss = 2.873540 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:25:51 INFO: Finished STEP 760/50000, loss = 3.564713 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:25:52 INFO: Finished STEP 780/50000, loss = 3.640228 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:25:53 INFO: Finished STEP 800/50000, loss = 2.930106 (0.036 sec/batch), lr: 0.003000
-2025-12-03 23:25:53 INFO: Evaluating on dev set...
-2025-12-03 23:25:53 INFO: LAS	MLAS	BLEX
-2025-12-03 23:25:53 INFO: 24.75	8.60	12.90
-2025-12-03 23:25:53 INFO: step 800: train_loss = 3.300487, dev_score = 0.2475
-2025-12-03 23:25:54 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:25:54 INFO: new model checkpoint saved.
-2025-12-03 23:25:54 INFO: Finished STEP 820/50000, loss = 2.968561 (0.035 sec/batch), lr: 0.003000
-2025-12-03 23:25:55 INFO: Finished STEP 840/50000, loss = 2.491823 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:25:56 INFO: Finished STEP 860/50000, loss = 3.991972 (0.035 sec/batch), lr: 0.003000
-2025-12-03 23:25:57 INFO: Finished STEP 880/50000, loss = 2.541115 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:25:58 INFO: Finished STEP 900/50000, loss = 3.015432 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:25:58 INFO: Evaluating on dev set...
-2025-12-03 23:25:58 INFO: LAS	MLAS	BLEX
-2025-12-03 23:25:58 INFO: 25.50	11.48	13.52
-2025-12-03 23:25:58 INFO: step 900: train_loss = 3.187365, dev_score = 0.2550
-2025-12-03 23:25:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:25:59 INFO: new model checkpoint saved.
-2025-12-03 23:25:59 INFO: Finished STEP 920/50000, loss = 2.031316 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:26:00 INFO: Finished STEP 940/50000, loss = 2.938839 (0.036 sec/batch), lr: 0.003000
-2025-12-03 23:26:01 INFO: Finished STEP 960/50000, loss = 3.606135 (0.036 sec/batch), lr: 0.003000
-2025-12-03 23:26:02 INFO: Finished STEP 980/50000, loss = 5.427132 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:26:02 INFO: Finished STEP 1000/50000, loss = 2.710342 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:26:02 INFO: Evaluating on dev set...
-2025-12-03 23:26:03 INFO: LAS	MLAS	BLEX
-2025-12-03 23:26:03 INFO: 37.87	21.71	24.22
-2025-12-03 23:26:03 INFO: step 1000: train_loss = 3.299336, dev_score = 0.3787
-2025-12-03 23:26:03 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:26:03 INFO: new best model saved.
-2025-12-03 23:26:04 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:26:04 INFO: new model checkpoint saved.
-2025-12-03 23:26:05 INFO: Finished STEP 1020/50000, loss = 3.274420 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:26:05 INFO: Finished STEP 1040/50000, loss = 3.172289 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:26:06 INFO: Finished STEP 1060/50000, loss = 2.884028 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:26:07 INFO: Finished STEP 1080/50000, loss = 4.205043 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:26:08 INFO: Finished STEP 1100/50000, loss = 3.608851 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:26:08 INFO: Evaluating on dev set...
-2025-12-03 23:26:08 INFO: LAS	MLAS	BLEX
-2025-12-03 23:26:08 INFO: 38.61	24.19	26.78
-2025-12-03 23:26:08 INFO: step 1100: train_loss = 3.198060, dev_score = 0.3861
-2025-12-03 23:26:09 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:26:09 INFO: new best model saved.
-2025-12-03 23:26:09 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:26:09 INFO: new model checkpoint saved.
-2025-12-03 23:26:10 INFO: Finished STEP 1120/50000, loss = 2.438182 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:26:11 INFO: Finished STEP 1140/50000, loss = 4.315670 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:26:11 INFO: Finished STEP 1160/50000, loss = 3.651882 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:26:12 INFO: Finished STEP 1180/50000, loss = 3.139420 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:26:13 INFO: Finished STEP 1200/50000, loss = 2.040530 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:26:13 INFO: Evaluating on dev set...
-2025-12-03 23:26:14 INFO: LAS	MLAS	BLEX
-2025-12-03 23:26:14 INFO: 39.85	23.98	26.98
-2025-12-03 23:26:14 INFO: step 1200: train_loss = 3.227944, dev_score = 0.3985
-2025-12-03 23:26:14 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:26:14 INFO: new best model saved.
-2025-12-03 23:26:14 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:26:14 INFO: new model checkpoint saved.
-2025-12-03 23:26:15 INFO: Finished STEP 1220/50000, loss = 2.697804 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:26:16 INFO: Finished STEP 1240/50000, loss = 2.911831 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:26:17 INFO: Finished STEP 1260/50000, loss = 1.945185 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:26:18 INFO: Finished STEP 1280/50000, loss = 3.025609 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:26:18 INFO: Finished STEP 1300/50000, loss = 4.230402 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:26:18 INFO: Evaluating on dev set...
-2025-12-03 23:26:19 INFO: LAS	MLAS	BLEX
-2025-12-03 23:26:19 INFO: 45.30	30.97	33.55
-2025-12-03 23:26:19 INFO: step 1300: train_loss = 3.111315, dev_score = 0.4530
-2025-12-03 23:26:19 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:26:19 INFO: new best model saved.
-2025-12-03 23:26:20 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:26:20 INFO: new model checkpoint saved.
-2025-12-03 23:26:21 INFO: Finished STEP 1320/50000, loss = 1.966291 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:26:21 INFO: Finished STEP 1340/50000, loss = 3.216881 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:26:22 INFO: Finished STEP 1360/50000, loss = 2.379959 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:26:23 INFO: Finished STEP 1380/50000, loss = 4.992296 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:26:24 INFO: Finished STEP 1400/50000, loss = 3.349003 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:26:24 INFO: Evaluating on dev set...
-2025-12-03 23:26:24 INFO: LAS	MLAS	BLEX
-2025-12-03 23:26:24 INFO: 36.88	23.21	24.89
-2025-12-03 23:26:24 INFO: step 1400: train_loss = 3.364135, dev_score = 0.3688
-2025-12-03 23:26:25 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:26:25 INFO: new model checkpoint saved.
-2025-12-03 23:26:26 INFO: Finished STEP 1420/50000, loss = 4.139513 (0.036 sec/batch), lr: 0.003000
-2025-12-03 23:26:26 INFO: Finished STEP 1440/50000, loss = 2.905265 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:26:27 INFO: Finished STEP 1460/50000, loss = 3.610150 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:26:28 INFO: Finished STEP 1480/50000, loss = 4.759534 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:26:29 INFO: Finished STEP 1500/50000, loss = 2.334270 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:26:29 INFO: Evaluating on dev set...
-2025-12-03 23:26:29 INFO: LAS	MLAS	BLEX
-2025-12-03 23:26:29 INFO: 33.66	20.92	32.22
-2025-12-03 23:26:29 INFO: step 1500: train_loss = 3.377781, dev_score = 0.3366
-2025-12-03 23:26:30 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:26:30 INFO: new model checkpoint saved.
-2025-12-03 23:26:31 INFO: Finished STEP 1520/50000, loss = 2.824460 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:26:32 INFO: Finished STEP 1540/50000, loss = 4.636815 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:26:32 INFO: Finished STEP 1560/50000, loss = 3.666232 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:26:33 INFO: Finished STEP 1580/50000, loss = 2.960546 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:26:34 INFO: Finished STEP 1600/50000, loss = 2.503325 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:26:34 INFO: Evaluating on dev set...
-2025-12-03 23:26:34 INFO: LAS	MLAS	BLEX
-2025-12-03 23:26:34 INFO: 42.57	27.91	31.29
-2025-12-03 23:26:34 INFO: step 1600: train_loss = 3.303970, dev_score = 0.4257
-2025-12-03 23:26:35 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:26:35 INFO: new model checkpoint saved.
-2025-12-03 23:26:36 INFO: Finished STEP 1620/50000, loss = 3.455626 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:26:37 INFO: Finished STEP 1640/50000, loss = 3.333741 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:26:37 INFO: Finished STEP 1660/50000, loss = 3.761724 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:26:38 INFO: Finished STEP 1680/50000, loss = 3.171466 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:26:39 INFO: Finished STEP 1700/50000, loss = 3.169466 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:26:39 INFO: Evaluating on dev set...
-2025-12-03 23:26:39 INFO: LAS	MLAS	BLEX
-2025-12-03 23:26:39 INFO: 45.05	33.96	39.41
-2025-12-03 23:26:39 INFO: step 1700: train_loss = 3.350673, dev_score = 0.4505
-2025-12-03 23:26:40 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:26:40 INFO: new model checkpoint saved.
-2025-12-03 23:26:41 INFO: Finished STEP 1720/50000, loss = 2.898433 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:26:42 INFO: Finished STEP 1740/50000, loss = 4.455222 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:26:42 INFO: Finished STEP 1760/50000, loss = 2.290793 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:26:43 INFO: Finished STEP 1780/50000, loss = 3.614108 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:26:44 INFO: Finished STEP 1800/50000, loss = 2.709010 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:26:44 INFO: Evaluating on dev set...
-2025-12-03 23:26:45 INFO: LAS	MLAS	BLEX
-2025-12-03 23:26:45 INFO: 46.04	29.96	33.76
-2025-12-03 23:26:45 INFO: step 1800: train_loss = 3.388469, dev_score = 0.4604
-2025-12-03 23:26:45 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:26:45 INFO: new best model saved.
-2025-12-03 23:26:45 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:26:45 INFO: new model checkpoint saved.
-2025-12-03 23:26:46 INFO: Finished STEP 1820/50000, loss = 3.027663 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:26:47 INFO: Finished STEP 1840/50000, loss = 4.950170 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:26:48 INFO: Finished STEP 1860/50000, loss = 3.627461 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:26:49 INFO: Finished STEP 1880/50000, loss = 3.900440 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:26:49 INFO: Finished STEP 1900/50000, loss = 3.206836 (0.037 sec/batch), lr: 0.003000
-2025-12-03 23:26:49 INFO: Evaluating on dev set...
-2025-12-03 23:26:50 INFO: LAS	MLAS	BLEX
-2025-12-03 23:26:50 INFO: 41.09	28.63	34.11
-2025-12-03 23:26:50 INFO: step 1900: train_loss = 3.372791, dev_score = 0.4109
-2025-12-03 23:26:51 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:26:51 INFO: new model checkpoint saved.
-2025-12-03 23:26:51 INFO: Finished STEP 1920/50000, loss = 4.990382 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:26:52 INFO: Finished STEP 1940/50000, loss = 3.309713 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:26:53 INFO: Finished STEP 1960/50000, loss = 3.610244 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:26:54 INFO: Finished STEP 1980/50000, loss = 3.427943 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:26:55 INFO: Finished STEP 2000/50000, loss = 3.286536 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:26:55 INFO: Evaluating on dev set...
-2025-12-03 23:26:55 INFO: LAS	MLAS	BLEX
-2025-12-03 23:26:55 INFO: 46.04	34.82	37.79
-2025-12-03 23:26:55 INFO: step 2000: train_loss = 3.315330, dev_score = 0.4604
-2025-12-03 23:26:55 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:26:55 INFO: new best model saved.
-2025-12-03 23:26:56 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:26:56 INFO: new model checkpoint saved.
-2025-12-03 23:26:57 INFO: Finished STEP 2020/50000, loss = 2.214467 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:26:58 INFO: Finished STEP 2040/50000, loss = 3.128998 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:26:58 INFO: Finished STEP 2060/50000, loss = 3.400111 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:26:59 INFO: Finished STEP 2080/50000, loss = 5.836899 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:27:00 INFO: Finished STEP 2100/50000, loss = 2.544196 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:27:00 INFO: Evaluating on dev set...
-2025-12-03 23:27:01 INFO: LAS	MLAS	BLEX
-2025-12-03 23:27:01 INFO: 41.34	27.60	36.09
-2025-12-03 23:27:01 INFO: step 2100: train_loss = 3.418321, dev_score = 0.4134
-2025-12-03 23:27:01 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:27:01 INFO: new model checkpoint saved.
-2025-12-03 23:27:02 INFO: Finished STEP 2120/50000, loss = 4.017645 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:27:03 INFO: Finished STEP 2140/50000, loss = 4.332951 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:27:04 INFO: Finished STEP 2160/50000, loss = 2.522452 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:27:04 INFO: Finished STEP 2180/50000, loss = 3.793148 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:27:05 INFO: Finished STEP 2200/50000, loss = 2.889349 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:27:05 INFO: Evaluating on dev set...
-2025-12-03 23:27:06 INFO: LAS	MLAS	BLEX
-2025-12-03 23:27:06 INFO: 39.60	26.64	35.10
-2025-12-03 23:27:06 INFO: step 2200: train_loss = 3.486956, dev_score = 0.3960
-2025-12-03 23:27:06 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:27:06 INFO: new model checkpoint saved.
-2025-12-03 23:27:07 INFO: Finished STEP 2220/50000, loss = 3.812532 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:27:08 INFO: Finished STEP 2240/50000, loss = 4.048740 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:27:09 INFO: Finished STEP 2260/50000, loss = 3.259020 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:27:10 INFO: Finished STEP 2280/50000, loss = 3.957003 (0.038 sec/batch), lr: 0.003000
-2025-12-03 23:27:10 INFO: Finished STEP 2300/50000, loss = 3.192188 (0.044 sec/batch), lr: 0.003000
-2025-12-03 23:27:10 INFO: Evaluating on dev set...
-2025-12-03 23:27:11 INFO: LAS	MLAS	BLEX
-2025-12-03 23:27:11 INFO: 43.81	27.85	33.33
-2025-12-03 23:27:11 INFO: step 2300: train_loss = 3.390584, dev_score = 0.4381
-2025-12-03 23:27:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:27:12 INFO: new model checkpoint saved.
-2025-12-03 23:27:12 INFO: Finished STEP 2320/50000, loss = 4.818910 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:27:13 INFO: Finished STEP 2340/50000, loss = 2.406329 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:27:14 INFO: Finished STEP 2360/50000, loss = 2.298516 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:27:15 INFO: Finished STEP 2380/50000, loss = 2.806631 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:27:16 INFO: Finished STEP 2400/50000, loss = 3.427664 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:27:16 INFO: Evaluating on dev set...
-2025-12-03 23:27:16 INFO: LAS	MLAS	BLEX
-2025-12-03 23:27:16 INFO: 43.07	31.09	37.39
-2025-12-03 23:27:16 INFO: step 2400: train_loss = 3.488748, dev_score = 0.4307
-2025-12-03 23:27:17 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:27:17 INFO: new model checkpoint saved.
-2025-12-03 23:27:18 INFO: Finished STEP 2420/50000, loss = 2.539723 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:27:19 INFO: Finished STEP 2440/50000, loss = 4.533563 (0.043 sec/batch), lr: 0.003000
-2025-12-03 23:27:19 INFO: Finished STEP 2460/50000, loss = 4.236969 (0.043 sec/batch), lr: 0.003000
-2025-12-03 23:27:20 INFO: Finished STEP 2480/50000, loss = 2.358161 (0.044 sec/batch), lr: 0.003000
-2025-12-03 23:27:21 INFO: Finished STEP 2500/50000, loss = 3.165836 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:27:21 INFO: Evaluating on dev set...
-2025-12-03 23:27:22 INFO: LAS	MLAS	BLEX
-2025-12-03 23:27:22 INFO: 46.53	30.44	35.52
-2025-12-03 23:27:22 INFO: step 2500: train_loss = 3.440427, dev_score = 0.4653
-2025-12-03 23:27:22 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:27:22 INFO: new best model saved.
-2025-12-03 23:27:22 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:27:22 INFO: new model checkpoint saved.
-2025-12-03 23:27:23 INFO: Finished STEP 2520/50000, loss = 3.732689 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:27:24 INFO: Finished STEP 2540/50000, loss = 5.190745 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:27:25 INFO: Finished STEP 2560/50000, loss = 4.239511 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:27:26 INFO: Finished STEP 2580/50000, loss = 4.053186 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:27:27 INFO: Finished STEP 2600/50000, loss = 2.809570 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:27:27 INFO: Evaluating on dev set...
-2025-12-03 23:27:27 INFO: LAS	MLAS	BLEX
-2025-12-03 23:27:27 INFO: 49.50	37.55	41.77
-2025-12-03 23:27:27 INFO: step 2600: train_loss = 3.507095, dev_score = 0.4950
-2025-12-03 23:27:27 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:27:27 INFO: new best model saved.
-2025-12-03 23:27:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:27:28 INFO: new model checkpoint saved.
-2025-12-03 23:27:29 INFO: Finished STEP 2620/50000, loss = 4.161043 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:27:30 INFO: Finished STEP 2640/50000, loss = 3.440454 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:27:31 INFO: Finished STEP 2660/50000, loss = 3.371374 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:27:31 INFO: Finished STEP 2680/50000, loss = 2.921980 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:27:32 INFO: Finished STEP 2700/50000, loss = 2.089964 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:27:32 INFO: Evaluating on dev set...
-2025-12-03 23:27:33 INFO: LAS	MLAS	BLEX
-2025-12-03 23:27:33 INFO: 45.54	33.12	36.94
-2025-12-03 23:27:33 INFO: step 2700: train_loss = 3.381599, dev_score = 0.4554
-2025-12-03 23:27:33 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:27:33 INFO: new model checkpoint saved.
-2025-12-03 23:27:34 INFO: Finished STEP 2720/50000, loss = 5.878531 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:27:35 INFO: Finished STEP 2740/50000, loss = 3.237665 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:27:36 INFO: Finished STEP 2760/50000, loss = 2.492691 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:27:37 INFO: Finished STEP 2780/50000, loss = 4.720194 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:27:38 INFO: Finished STEP 2800/50000, loss = 3.760880 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:27:38 INFO: Evaluating on dev set...
-2025-12-03 23:27:38 INFO: LAS	MLAS	BLEX
-2025-12-03 23:27:38 INFO: 51.24	38.24	43.28
-2025-12-03 23:27:38 INFO: step 2800: train_loss = 3.545646, dev_score = 0.5124
-2025-12-03 23:27:38 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:27:38 INFO: new best model saved.
-2025-12-03 23:27:39 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:27:39 INFO: new model checkpoint saved.
-2025-12-03 23:27:40 INFO: Finished STEP 2820/50000, loss = 3.466887 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:27:41 INFO: Finished STEP 2840/50000, loss = 2.830301 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:27:42 INFO: Finished STEP 2860/50000, loss = 3.183891 (0.043 sec/batch), lr: 0.003000
-2025-12-03 23:27:43 INFO: Finished STEP 2880/50000, loss = 3.444857 (0.043 sec/batch), lr: 0.003000
-2025-12-03 23:27:43 INFO: Finished STEP 2900/50000, loss = 3.642260 (0.044 sec/batch), lr: 0.003000
-2025-12-03 23:27:43 INFO: Evaluating on dev set...
-2025-12-03 23:27:44 INFO: LAS	MLAS	BLEX
-2025-12-03 23:27:44 INFO: 53.71	40.59	44.82
-2025-12-03 23:27:44 INFO: step 2900: train_loss = 3.663115, dev_score = 0.5371
-2025-12-03 23:27:44 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:27:44 INFO: new best model saved.
-2025-12-03 23:27:45 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:27:45 INFO: new model checkpoint saved.
-2025-12-03 23:27:46 INFO: Finished STEP 2920/50000, loss = 2.488231 (0.043 sec/batch), lr: 0.003000
-2025-12-03 23:27:46 INFO: Finished STEP 2940/50000, loss = 3.432896 (0.044 sec/batch), lr: 0.003000
-2025-12-03 23:27:47 INFO: Finished STEP 2960/50000, loss = 2.781652 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:27:48 INFO: Finished STEP 2980/50000, loss = 2.840132 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:27:49 INFO: Finished STEP 3000/50000, loss = 4.171174 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:27:49 INFO: Evaluating on dev set...
-2025-12-03 23:27:49 INFO: LAS	MLAS	BLEX
-2025-12-03 23:27:49 INFO: 51.98	39.50	45.80
-2025-12-03 23:27:49 INFO: step 3000: train_loss = 3.476293, dev_score = 0.5198
-2025-12-03 23:27:50 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:27:50 INFO: new model checkpoint saved.
-2025-12-03 23:27:51 INFO: Finished STEP 3020/50000, loss = 3.225040 (0.043 sec/batch), lr: 0.003000
-2025-12-03 23:27:52 INFO: Finished STEP 3040/50000, loss = 3.064417 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:27:53 INFO: Finished STEP 3060/50000, loss = 3.886524 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:27:53 INFO: Finished STEP 3080/50000, loss = 3.704923 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:27:54 INFO: Finished STEP 3100/50000, loss = 4.191244 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:27:54 INFO: Evaluating on dev set...
-2025-12-03 23:27:55 INFO: LAS	MLAS	BLEX
-2025-12-03 23:27:55 INFO: 49.75	37.32	40.67
-2025-12-03 23:27:55 INFO: step 3100: train_loss = 3.466918, dev_score = 0.4975
-2025-12-03 23:27:55 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:27:55 INFO: new model checkpoint saved.
-2025-12-03 23:27:56 INFO: Finished STEP 3120/50000, loss = 3.950969 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:27:57 INFO: Finished STEP 3140/50000, loss = 2.399348 (0.039 sec/batch), lr: 0.003000
-2025-12-03 23:27:58 INFO: Finished STEP 3160/50000, loss = 3.456824 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:27:59 INFO: Finished STEP 3180/50000, loss = 3.288764 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:28:00 INFO: Finished STEP 3200/50000, loss = 3.057923 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:28:00 INFO: Evaluating on dev set...
-2025-12-03 23:28:00 INFO: LAS	MLAS	BLEX
-2025-12-03 23:28:00 INFO: 47.77	36.55	40.76
-2025-12-03 23:28:00 INFO: step 3200: train_loss = 3.563295, dev_score = 0.4777
-2025-12-03 23:28:01 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:28:01 INFO: new model checkpoint saved.
-2025-12-03 23:28:02 INFO: Finished STEP 3220/50000, loss = 4.334808 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:28:03 INFO: Finished STEP 3240/50000, loss = 2.773743 (0.040 sec/batch), lr: 0.003000
-2025-12-03 23:28:03 INFO: Finished STEP 3260/50000, loss = 3.285001 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:28:04 INFO: Finished STEP 3280/50000, loss = 3.142590 (0.046 sec/batch), lr: 0.003000
-2025-12-03 23:28:05 INFO: Finished STEP 3300/50000, loss = 2.988616 (0.044 sec/batch), lr: 0.003000
-2025-12-03 23:28:05 INFO: Evaluating on dev set...
-2025-12-03 23:28:06 INFO: LAS	MLAS	BLEX
-2025-12-03 23:28:06 INFO: 51.98	39.08	41.60
-2025-12-03 23:28:06 INFO: step 3300: train_loss = 3.589203, dev_score = 0.5198
-2025-12-03 23:28:06 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:28:06 INFO: new model checkpoint saved.
-2025-12-03 23:28:07 INFO: Finished STEP 3320/50000, loss = 5.763182 (0.045 sec/batch), lr: 0.003000
-2025-12-03 23:28:08 INFO: Finished STEP 3340/50000, loss = 3.421094 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:28:09 INFO: Finished STEP 3360/50000, loss = 5.406409 (0.043 sec/batch), lr: 0.003000
-2025-12-03 23:28:10 INFO: Finished STEP 3380/50000, loss = 3.032209 (0.045 sec/batch), lr: 0.003000
-2025-12-03 23:28:11 INFO: Finished STEP 3400/50000, loss = 3.139112 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:28:11 INFO: Evaluating on dev set...
-2025-12-03 23:28:11 INFO: LAS	MLAS	BLEX
-2025-12-03 23:28:11 INFO: 49.75	36.29	40.08
-2025-12-03 23:28:11 INFO: step 3400: train_loss = 3.722110, dev_score = 0.4975
-2025-12-03 23:28:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:28:12 INFO: new model checkpoint saved.
-2025-12-03 23:28:13 INFO: Finished STEP 3420/50000, loss = 3.515473 (0.044 sec/batch), lr: 0.003000
-2025-12-03 23:28:14 INFO: Finished STEP 3440/50000, loss = 4.545405 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:28:15 INFO: Finished STEP 3460/50000, loss = 3.489767 (0.043 sec/batch), lr: 0.003000
-2025-12-03 23:28:15 INFO: Finished STEP 3480/50000, loss = 4.931797 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:28:16 INFO: Finished STEP 3500/50000, loss = 3.305768 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:28:16 INFO: Evaluating on dev set...
-2025-12-03 23:28:17 INFO: LAS	MLAS	BLEX
-2025-12-03 23:28:17 INFO: 48.76	37.05	41.68
-2025-12-03 23:28:17 INFO: step 3500: train_loss = 3.662503, dev_score = 0.4876
-2025-12-03 23:28:17 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:28:17 INFO: new model checkpoint saved.
-2025-12-03 23:28:18 INFO: Finished STEP 3520/50000, loss = 4.402299 (0.043 sec/batch), lr: 0.003000
-2025-12-03 23:28:19 INFO: Finished STEP 3540/50000, loss = 2.635880 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:28:20 INFO: Finished STEP 3560/50000, loss = 2.857255 (0.043 sec/batch), lr: 0.003000
-2025-12-03 23:28:21 INFO: Finished STEP 3580/50000, loss = 3.507267 (0.043 sec/batch), lr: 0.003000
-2025-12-03 23:28:22 INFO: Finished STEP 3600/50000, loss = 5.196735 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:28:22 INFO: Evaluating on dev set...
-2025-12-03 23:28:22 INFO: LAS	MLAS	BLEX
-2025-12-03 23:28:22 INFO: 51.24	40.84	44.21
-2025-12-03 23:28:22 INFO: step 3600: train_loss = 3.681615, dev_score = 0.5124
-2025-12-03 23:28:23 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:28:23 INFO: new model checkpoint saved.
-2025-12-03 23:28:24 INFO: Finished STEP 3620/50000, loss = 4.539114 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:28:25 INFO: Finished STEP 3640/50000, loss = 6.413163 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:28:25 INFO: Finished STEP 3660/50000, loss = 4.375307 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:28:26 INFO: Finished STEP 3680/50000, loss = 5.120213 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:28:27 INFO: Finished STEP 3700/50000, loss = 3.109243 (0.044 sec/batch), lr: 0.003000
-2025-12-03 23:28:27 INFO: Evaluating on dev set...
-2025-12-03 23:28:28 INFO: LAS	MLAS	BLEX
-2025-12-03 23:28:28 INFO: 49.26	38.99	44.03
-2025-12-03 23:28:28 INFO: step 3700: train_loss = 3.778901, dev_score = 0.4926
-2025-12-03 23:28:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:28:28 INFO: new model checkpoint saved.
-2025-12-03 23:28:29 INFO: Finished STEP 3720/50000, loss = 3.468835 (0.045 sec/batch), lr: 0.003000
-2025-12-03 23:28:30 INFO: Finished STEP 3740/50000, loss = 3.420460 (0.044 sec/batch), lr: 0.003000
-2025-12-03 23:28:31 INFO: Finished STEP 3760/50000, loss = 2.856275 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:28:32 INFO: Finished STEP 3780/50000, loss = 2.668692 (0.041 sec/batch), lr: 0.003000
-2025-12-03 23:28:33 INFO: Finished STEP 3800/50000, loss = 4.511623 (0.044 sec/batch), lr: 0.003000
-2025-12-03 23:28:33 INFO: Evaluating on dev set...
-2025-12-03 23:28:33 INFO: LAS	MLAS	BLEX
-2025-12-03 23:28:33 INFO: 45.30	35.79	39.16
-2025-12-03 23:28:33 INFO: step 3800: train_loss = 3.920723, dev_score = 0.4530
-2025-12-03 23:28:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 23:28:34 INFO: new model checkpoint saved.
-2025-12-03 23:28:35 INFO: Finished STEP 3820/50000, loss = 4.449274 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:28:36 INFO: Finished STEP 3840/50000, loss = 2.533618 (0.042 sec/batch), lr: 0.003000
-2025-12-03 23:28:36 INFO: Finished STEP 3860/50000, loss = 3.834488 (0.043 sec/batch), lr: 0.003000
-2025-12-03 23:28:37 INFO: Finished STEP 3880/50000, loss = 3.449773 (0.043 sec/batch), lr: 0.003000
-2025-12-03 23:28:38 INFO: Finished STEP 3900/50000, loss = 3.354862 (0.047 sec/batch), lr: 0.003000
-2025-12-03 23:28:38 INFO: Evaluating on dev set...
-2025-12-03 23:28:39 INFO: LAS	MLAS	BLEX
-2025-12-03 23:28:39 INFO: 52.48	39.58	44.63
-2025-12-03 23:28:39 INFO: step 3900: train_loss = 3.766196, dev_score = 0.5248
-2025-12-03 23:28:39 INFO: Training ended with 3900 steps.
-2025-12-03 23:28:39 INFO: Best dev F1 = 53.71, at iteration = 2900
-2025-12-03 23:28:39 INFO: Running dev depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
-2025-12-03 23:28:39 INFO: Running parser in predict mode
-2025-12-03 23:28:39 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:28:41 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 23:28:41 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 23:28:41 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 23:28:41 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 23:28:41 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 23:28:41 INFO: Loading data with batch size 32...
-2025-12-03 23:28:41 DEBUG: 9 batches created.
-2025-12-03 23:28:42 INFO: F1 scores for each dependency:
-  Note that unlabeled attachment errors hurt the labeled attachment scores
-       acl: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
- acl:relcl: p 0.1667 r 0.2857 f1 0.2105 (7 actual)
-     advcl: p 0.0526 r 0.2000 f1 0.0833 (5 actual)
-    advmod: p 0.4839 r 0.6000 f1 0.5357 (25 actual)
-      amod: p 0.8148 r 0.7097 f1 0.7586 (31 actual)
-     appos: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-       aux: p 0.8889 r 0.7273 f1 0.8000 (11 actual)
-      case: p 0.7544 r 0.7679 f1 0.7611 (56 actual)
-        cc: p 0.7692 r 0.7692 f1 0.7692 (13 actual)
-     ccomp: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-      conj: p 0.5000 r 0.0833 f1 0.1429 (12 actual)
-       cop: p 0.2500 r 0.3333 f1 0.2857 (3 actual)
-     csubj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-       det: p 0.8696 r 0.9091 f1 0.8889 (22 actual)
-      expl: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-      iobj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-      mark: p 0.5000 r 0.3333 f1 0.4000 (12 actual)
-      nmod: p 0.2609 r 0.4000 f1 0.3158 (15 actual)
- nmod:poss: p 1.0000 r 0.8947 f1 0.9444 (19 actual)
-     nsubj: p 0.3421 r 0.7647 f1 0.4727 (17 actual)
-nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
-       obj: p 0.6667 r 0.2727 f1 0.3871 (22 actual)
-       obl: p 0.3333 r 0.4146 f1 0.3696 (41 actual)
- obl:agent: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-    orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
- parataxis: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
-     punct: p 0.4808 r 0.4808 f1 0.4808 (52 actual)
-      root: p 0.4444 r 0.4444 f1 0.4444 (9 actual)
-     xcomp: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
-2025-12-03 23:28:42 INFO: LAS	MLAS	BLEX
-2025-12-03 23:28:42 INFO: 53.71	40.59	44.82
-2025-12-03 23:28:42 INFO: Parser score:
-2025-12-03 23:28:42 INFO: sv_diachronic 53.71
-2025-12-03 23:28:42 INFO: Finished running dev set on
-UD_Swedish-diachronic
-  UAS   LAS  CLAS  MLAS  BLEX
-66.58 53.71 44.82 40.59 44.82
-2025-12-03 23:28:42 INFO: Running test depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
-2025-12-03 23:28:42 INFO: Running parser in predict mode
-2025-12-03 23:28:42 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 23:28:44 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 23:28:44 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 23:28:44 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 23:28:44 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 23:28:44 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 23:28:44 INFO: Loading data with batch size 32...
-2025-12-03 23:28:44 DEBUG: 93 batches created.
-2025-12-03 23:28:49 INFO: F1 scores for each dependency:
-  Note that unlabeled attachment errors hurt the labeled attachment scores
-         acl: p 0.0000 r 0.0000 f1 0.0000 (32 actual)
-   acl:cleft: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-   acl:relcl: p 0.1167 r 0.0933 f1 0.1037 (75 actual)
-       advcl: p 0.0544 r 0.2667 f1 0.0904 (60 actual)
- advcl:relcl: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-      advmod: p 0.5083 r 0.5746 f1 0.5394 (268 actual)
-        amod: p 0.6795 r 0.6913 f1 0.6853 (230 actual)
-       appos: p 0.0000 r 0.0000 f1 0.0000 (13 actual)
-         aux: p 0.7556 r 0.8095 f1 0.7816 (84 actual)
-    aux:pass: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-        case: p 0.7245 r 0.7051 f1 0.7147 (373 actual)
-          cc: p 0.5759 r 0.5871 f1 0.5815 (155 actual)
-       ccomp: p 0.0000 r 0.0000 f1 0.0000 (35 actual)
-compound:prt: p 0.0000 r 0.0000 f1 0.0000 (21 actual)
-        conj: p 0.1053 r 0.0253 f1 0.0408 (158 actual)
-         cop: p 0.7619 r 0.3478 f1 0.4776 (46 actual)
-       csubj: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-         dep: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-         det: p 0.7833 r 0.7644 f1 0.7737 (208 actual)
-   discourse: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
-  dislocated: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-        expl: p 0.0000 r 0.0000 f1 0.0000 (11 actual)
-     expl:pv: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-       fixed: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
-        flat: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-   flat:name: p 0.0000 r 0.0000 f1 0.0000 (12 actual)
-    goeswith: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-        iobj: p 0.0000 r 0.0000 f1 0.0000 (14 actual)
-        mark: p 0.6569 r 0.5882 f1 0.6207 (153 actual)
-        nmod: p 0.2320 r 0.2843 f1 0.2555 (102 actual)
-   nmod:poss: p 0.8102 r 0.7817 f1 0.7957 (142 actual)
-       nsubj: p 0.3947 r 0.6429 f1 0.4891 (280 actual)
-  nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (25 actual)
-      nummod: p 0.0000 r 0.0000 f1 0.0000 (10 actual)
-         obj: p 0.6111 r 0.1803 f1 0.2785 (183 actual)
-         obl: p 0.2623 r 0.4029 f1 0.3177 (278 actual)
-   obl:agent: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-      orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-   parataxis: p 0.0000 r 0.0000 f1 0.0000 (18 actual)
-       punct: p 0.3852 r 0.3906 f1 0.3879 (425 actual)
-  reparandum: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-        root: p 0.5253 r 0.5253 f1 0.5253 (99 actual)
-    vocative: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
-       xcomp: p 0.0000 r 0.0000 f1 0.0000 (75 actual)
-2025-12-03 23:28:49 INFO: LAS	MLAS	BLEX
-2025-12-03 23:28:49 INFO: 47.60	36.14	39.67
-2025-12-03 23:28:49 INFO: Parser score:
-2025-12-03 23:28:49 INFO: sv_diachronic 47.60
-2025-12-03 23:28:49 INFO: Finished running test set on
-UD_Swedish-diachronic
-  UAS   LAS  CLAS  MLAS  BLEX
-62.40 47.60 39.67 36.14 39.67
-DONE.
-Full log saved to: logs/log_conll17.pt_sv_20251203_232257.txt
-Symlink updated: logs/latest.txt → log_conll17.pt_sv_20251203_232257.txt

logs/log_conll17.pt_sv_20260218_190034.txt

@@ -1,120 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_20260218_190034.txt ===
-Language codes: sv
-Using pretrained model: conll17.pt
-
-Running: python prepare-train-val-SILVER.py sv
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Loading DigPhil MACHINE for SILVER (minus gold overlap)…
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
-Adding 75 remaining gold sentences to TRAIN
-Cleaning TRAIN...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Cleaning SILVER...
-[REMOVED] sent_id=11_12_13  ERRORS=['Multiple roots found: [5, 17, 25]']
-[REMOVED] sent_id=65_split2_66_split2  ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
-[REMOVED] sent_id=50_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
-[REMOVED] sent_id=53_54  ERRORS=['Multiple roots found: [27, 91]']
-[REMOVED] sent_id=55_56_57  ERRORS=['Multiple roots found: [2, 4, 13]']
-[REMOVED] sent_id=909  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=912  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=396  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=416  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=589  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=17_split1  ERRORS=['Multiple roots found: [2, 14, 17]']
-[REMOVED] sent_id=17_split2  ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
-[REMOVED] sent_id=19_split1  ERRORS=['Multiple roots found: [3, 31]']
-[REMOVED] sent_id=87_88  ERRORS=['Multiple roots found: [3, 6]']
-[REMOVED] sent_id=25  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=136  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=208  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=230  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=245  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=276  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=320  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=366  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=519  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=569  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=6  ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
-[REMOVED] sent_id=7_8  ERRORS=['Multiple roots found: [5, 10]']
-[REMOVED] sent_id=124_split2  ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
-[REMOVED] sent_id=30_31  ERRORS=['Multiple roots found: [3, 18]']
-[REMOVED] sent_id=35  ERRORS=['Line 36: Invalid token ID or head']
-[REMOVED] sent_id=119  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=188  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=223  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=268  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=325  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=388  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=399  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=475  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=505  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=520  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=562  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=711  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=731  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=867  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=884  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=923  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=939  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1086  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1251  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1345  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1459  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1656  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=3_split1  ERRORS=['Multiple roots found: [4, 15, 17]']
-[REMOVED] sent_id=3_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=3_4  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=5_6  ERRORS=['Multiple roots found: [3, 24]']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [7, 20]']
-[REMOVED] sent_id=8_9  ERRORS=['Multiple roots found: [24, 57]']
-[REMOVED] sent_id=12_13  ERRORS=['Multiple roots found: [11, 16]']
-Cleaning DEV...
-Writing TRAIN → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (13868 valid sentences)
-Writing SILVER → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu (46432 valid sentences)
-Writing DEV → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (34 valid sentences)
-Done.
-Sourcing scripts/config.sh
-Running stanza dataset preparation…
-2026-02-18 19:00:38 INFO: Datasets program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --use_bert --bert_model vesteinn/ScandiBERT --silver_file ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu
-usage: python -m stanza.utils.datasets.prepare_depparse_treebank
-       [-h] [--gold] [--predicted]
-       [--wordvec_pretrain_file WORDVEC_PRETRAIN_FILE]
-       [--tagger_model TAGGER_MODEL] [--save_dir SAVE_DIR]
-       [--no_download_tagger] [--charlm CHARLM] [--no_charlm]
-       treebanks [treebanks ...]
-python -m stanza.utils.datasets.prepare_depparse_treebank: error: unrecognized arguments: --use_bert --bert_model vesteinn/ScandiBERT --silver_file ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu

logs/log_conll17.pt_sv_20260218_190242.txt

@@ -1,143 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_20260218_190242.txt ===
-Language codes: sv
-Using pretrained model: conll17.pt
-
-Running: python prepare-train-val-SILVER.py sv
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Loading DigPhil MACHINE for SILVER (minus gold overlap)…
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
-Adding 75 remaining gold sentences to TRAIN
-Cleaning TRAIN...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Cleaning SILVER...
-[REMOVED] sent_id=11_12_13  ERRORS=['Multiple roots found: [5, 17, 25]']
-[REMOVED] sent_id=65_split2_66_split2  ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
-[REMOVED] sent_id=50_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
-[REMOVED] sent_id=53_54  ERRORS=['Multiple roots found: [27, 91]']
-[REMOVED] sent_id=55_56_57  ERRORS=['Multiple roots found: [2, 4, 13]']
-[REMOVED] sent_id=909  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=912  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=396  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=416  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=589  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=17_split1  ERRORS=['Multiple roots found: [2, 14, 17]']
-[REMOVED] sent_id=17_split2  ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
-[REMOVED] sent_id=19_split1  ERRORS=['Multiple roots found: [3, 31]']
-[REMOVED] sent_id=87_88  ERRORS=['Multiple roots found: [3, 6]']
-[REMOVED] sent_id=25  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=136  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=208  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=230  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=245  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=276  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=320  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=366  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=519  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=569  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=6  ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
-[REMOVED] sent_id=7_8  ERRORS=['Multiple roots found: [5, 10]']
-[REMOVED] sent_id=124_split2  ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
-[REMOVED] sent_id=30_31  ERRORS=['Multiple roots found: [3, 18]']
-[REMOVED] sent_id=35  ERRORS=['Line 36: Invalid token ID or head']
-[REMOVED] sent_id=119  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=188  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=223  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=268  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=325  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=388  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=399  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=475  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=505  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=520  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=562  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=711  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=731  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=867  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=884  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=923  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=939  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1086  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1251  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1345  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1459  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1656  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=3_split1  ERRORS=['Multiple roots found: [4, 15, 17]']
-[REMOVED] sent_id=3_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=3_4  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=5_6  ERRORS=['Multiple roots found: [3, 24]']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [7, 20]']
-[REMOVED] sent_id=8_9  ERRORS=['Multiple roots found: [24, 57]']
-[REMOVED] sent_id=12_13  ERRORS=['Multiple roots found: [11, 16]']
-Cleaning DEV...
-Writing TRAIN → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (13868 valid sentences)
-Writing SILVER → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu (46432 valid sentences)
-Writing DEV → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (34 valid sentences)
-Done.
-Sourcing scripts/config.sh
-Running stanza dataset preparation…
-2026-02-18 19:02:45 INFO: Datasets program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic
-2026-02-18 19:02:45 DEBUG: Looking for tagger for lang |sv| dataset |diachronic|.  Suggested model |None|.  Looking first in |saved_models/pos|.
-2026-02-18 19:02:45 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
-
-2026-02-18 19:02:46 INFO: Downloaded file to /home/urdatorn/stanza_resources/resources.json
-2026-02-18 19:02:46 DEBUG: Processing parameter "processors"...
-2026-02-18 19:02:46 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
-2026-02-18 19:02:46 INFO: Downloading these customized packages for language: sv (Swedish)...
-=======================
-| Processor | Package |
------------------------
-=======================
-
-2026-02-18 19:02:46 INFO: Finished downloading models and saved to /home/urdatorn/stanza_resources
-Traceback (most recent call last):
-  File "<frozen runpy>", line 198, in _run_module_as_main
-  File "<frozen runpy>", line 88, in _run_code
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 151, in <module>
-    main()
-    ~~~~^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 147, in main
-    common.main(process_treebank, common.ModelType.DEPPARSE, add_specific_args)
-    ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/common.py", line 335, in main
-    process_treebank(treebank, model_type, paths, args)
-    ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 109, in process_treebank
-    tagger_model = choose_tagger_model(short_language, dataset, args.tagger_model, args)
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 90, in choose_tagger_model
-    raise FileNotFoundError("Could not figure out which model file to use for %s.  Just tried to download to %s the models %s" % (short_language, args.save_dir, download_list))
-FileNotFoundError: Could not figure out which model file to use for sv.  Just tried to download to saved_models/pos the models []

logs/log_conll17.pt_sv_20260218_190400.txt

@@ -1,38 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_20260218_190400.txt ===
-Language codes: sv
-Using pretrained model: conll17.pt
-
-Sourcing scripts/config.sh
-Running stanza dataset preparation…
-2026-02-18 19:04:02 INFO: Datasets program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic
-2026-02-18 19:04:02 DEBUG: Looking for tagger for lang |sv| dataset |diachronic|.  Suggested model |None|.  Looking first in |saved_models/pos|.
-2026-02-18 19:04:02 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
-
-2026-02-18 19:04:02 INFO: Downloaded file to /home/urdatorn/stanza_resources/resources.json
-2026-02-18 19:04:02 DEBUG: Processing parameter "processors"...
-2026-02-18 19:04:02 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
-2026-02-18 19:04:02 INFO: Downloading these customized packages for language: sv (Swedish)...
-=======================
-| Processor | Package |
------------------------
-=======================
-
-2026-02-18 19:04:02 INFO: Finished downloading models and saved to /home/urdatorn/stanza_resources
-Traceback (most recent call last):
-  File "<frozen runpy>", line 198, in _run_module_as_main
-  File "<frozen runpy>", line 88, in _run_code
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 151, in <module>
-    main()
-    ~~~~^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 147, in main
-    common.main(process_treebank, common.ModelType.DEPPARSE, add_specific_args)
-    ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/common.py", line 335, in main
-    process_treebank(treebank, model_type, paths, args)
-    ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 109, in process_treebank
-    tagger_model = choose_tagger_model(short_language, dataset, args.tagger_model, args)
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 90, in choose_tagger_model
-    raise FileNotFoundError("Could not figure out which model file to use for %s.  Just tried to download to %s the models %s" % (short_language, args.save_dir, download_list))
-FileNotFoundError: Could not figure out which model file to use for sv.  Just tried to download to saved_models/pos the models []

logs/log_conll17.pt_sv_20260218_190808.txt

@@ -1,38 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_20260218_190808.txt ===
-Language codes: sv
-Using pretrained model: conll17.pt
-
-Sourcing scripts/config.sh
-Running stanza dataset preparation…
-2026-02-18 19:08:10 INFO: Datasets program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic
-2026-02-18 19:08:10 DEBUG: Looking for tagger for lang |sv| dataset |diachronic|.  Suggested model |None|.  Looking first in |saved_models/pos|.
-2026-02-18 19:08:10 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
-
-2026-02-18 19:08:11 INFO: Downloaded file to /home/urdatorn/stanza_resources/resources.json
-2026-02-18 19:08:11 DEBUG: Processing parameter "processors"...
-2026-02-18 19:08:11 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
-2026-02-18 19:08:11 INFO: Downloading these customized packages for language: sv (Swedish)...
-=======================
-| Processor | Package |
------------------------
-=======================
-
-2026-02-18 19:08:11 INFO: Finished downloading models and saved to /home/urdatorn/stanza_resources
-Traceback (most recent call last):
-  File "<frozen runpy>", line 198, in _run_module_as_main
-  File "<frozen runpy>", line 88, in _run_code
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 151, in <module>
-    main()
-    ~~~~^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 147, in main
-    common.main(process_treebank, common.ModelType.DEPPARSE, add_specific_args)
-    ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/common.py", line 335, in main
-    process_treebank(treebank, model_type, paths, args)
-    ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 109, in process_treebank
-    tagger_model = choose_tagger_model(short_language, dataset, args.tagger_model, args)
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 90, in choose_tagger_model
-    raise FileNotFoundError("Could not figure out which model file to use for %s.  Just tried to download to %s the models %s" % (short_language, args.save_dir, download_list))
-FileNotFoundError: Could not figure out which model file to use for sv.  Just tried to download to saved_models/pos the models []

logs/log_conll17.pt_sv_20260218_190956.txt

@@ -1,38 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_20260218_190956.txt ===
-Language codes: sv
-Using pretrained model: conll17.pt
-
-Sourcing scripts/config.sh
-Running stanza dataset preparation…
-2026-02-18 19:09:58 INFO: Datasets program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt
-2026-02-18 19:09:58 DEBUG: Looking for tagger for lang |sv| dataset |diachronic|.  Suggested model |None|.  Looking first in |saved_models/pos|.
-2026-02-18 19:09:58 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
-
-2026-02-18 19:09:58 INFO: Downloaded file to /home/urdatorn/stanza_resources/resources.json
-2026-02-18 19:09:58 DEBUG: Processing parameter "processors"...
-2026-02-18 19:09:58 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
-2026-02-18 19:09:58 INFO: Downloading these customized packages for language: sv (Swedish)...
-=======================
-| Processor | Package |
------------------------
-=======================
-
-2026-02-18 19:09:58 INFO: Finished downloading models and saved to /home/urdatorn/stanza_resources
-Traceback (most recent call last):
-  File "<frozen runpy>", line 198, in _run_module_as_main
-  File "<frozen runpy>", line 88, in _run_code
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 151, in <module>
-    main()
-    ~~~~^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 147, in main
-    common.main(process_treebank, common.ModelType.DEPPARSE, add_specific_args)
-    ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/common.py", line 335, in main
-    process_treebank(treebank, model_type, paths, args)
-    ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 109, in process_treebank
-    tagger_model = choose_tagger_model(short_language, dataset, args.tagger_model, args)
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 90, in choose_tagger_model
-    raise FileNotFoundError("Could not figure out which model file to use for %s.  Just tried to download to %s the models %s" % (short_language, args.save_dir, download_list))
-FileNotFoundError: Could not figure out which model file to use for sv.  Just tried to download to saved_models/pos the models []

logs/log_conll17.pt_sv_20260218_191305.txt

@@ -1,38 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_20260218_191305.txt ===
-Language codes: sv
-Using pretrained model: conll17.pt
-
-Sourcing scripts/config.sh
-Running stanza dataset preparation…
-2026-02-18 19:13:07 INFO: Datasets program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt
-2026-02-18 19:13:07 DEBUG: Looking for tagger for lang |sv| dataset |diachronic|.  Suggested model |None|.  Looking first in |saved_models/pos|.
-2026-02-18 19:13:07 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
-
-2026-02-18 19:13:07 INFO: Downloaded file to /home/urdatorn/stanza_resources/resources.json
-2026-02-18 19:13:07 DEBUG: Processing parameter "processors"...
-2026-02-18 19:13:07 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
-2026-02-18 19:13:07 INFO: Downloading these customized packages for language: sv (Swedish)...
-=======================
-| Processor | Package |
------------------------
-=======================
-
-2026-02-18 19:13:07 INFO: Finished downloading models and saved to /home/urdatorn/stanza_resources
-Traceback (most recent call last):
-  File "<frozen runpy>", line 198, in _run_module_as_main
-  File "<frozen runpy>", line 88, in _run_code
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 151, in <module>
-    main()
-    ~~~~^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 147, in main
-    common.main(process_treebank, common.ModelType.DEPPARSE, add_specific_args)
-    ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/common.py", line 335, in main
-    process_treebank(treebank, model_type, paths, args)
-    ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 109, in process_treebank
-    tagger_model = choose_tagger_model(short_language, dataset, args.tagger_model, args)
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 90, in choose_tagger_model
-    raise FileNotFoundError("Could not figure out which model file to use for %s.  Just tried to download to %s the models %s" % (short_language, args.save_dir, download_list))
-FileNotFoundError: Could not figure out which model file to use for sv.  Just tried to download to saved_models/pos the models []

logs/log_conll17.pt_sv_20260218_191719.txt

@@ -1,38 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_20260218_191719.txt ===
-Language codes: sv
-Using pretrained model: conll17.pt
-
-Sourcing scripts/config.sh
-Running stanza dataset preparation…
-2026-02-18 19:17:21 INFO: Datasets program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt
-2026-02-18 19:17:21 DEBUG: Looking for tagger for lang |sv| dataset |diachronic|.  Suggested model |None|.  Looking first in |saved_models/pos|.
-2026-02-18 19:17:21 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
-
-2026-02-18 19:17:21 INFO: Downloaded file to /home/urdatorn/stanza_resources/resources.json
-2026-02-18 19:17:21 DEBUG: Processing parameter "processors"...
-2026-02-18 19:17:21 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
-2026-02-18 19:17:21 INFO: Downloading these customized packages for language: sv (Swedish)...
-=======================
-| Processor | Package |
------------------------
-=======================
-
-2026-02-18 19:17:21 INFO: Finished downloading models and saved to /home/urdatorn/stanza_resources
-Traceback (most recent call last):
-  File "<frozen runpy>", line 198, in _run_module_as_main
-  File "<frozen runpy>", line 88, in _run_code
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 151, in <module>
-    main()
-    ~~~~^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 147, in main
-    common.main(process_treebank, common.ModelType.DEPPARSE, add_specific_args)
-    ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/common.py", line 335, in main
-    process_treebank(treebank, model_type, paths, args)
-    ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 109, in process_treebank
-    tagger_model = choose_tagger_model(short_language, dataset, args.tagger_model, args)
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 90, in choose_tagger_model
-    raise FileNotFoundError("Could not figure out which model file to use for %s.  Just tried to download to %s the models %s" % (short_language, args.save_dir, download_list))
-FileNotFoundError: Could not figure out which model file to use for sv.  Just tried to download to saved_models/pos the models []

logs/log_conll17.pt_sv_20260219_104841.txt

@@ -1,44 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_20260219_104841.txt ===
-Language codes: sv
-Using pretrained model: conll17.pt
-
-Sourcing scripts/config.sh
-Running stanza dataset preparation…
-2026-02-19 10:48:43 INFO: Datasets program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt --gold
-Augmented 60 quotes: Counter({'《》': 9, '»«': 8, '″″': 8, '““': 7, '„”': 6, '""': 6, '„“': 6, '«»': 4, '””': 4, '「」': 2})
-Traceback (most recent call last):
-  File "<frozen runpy>", line 198, in _run_module_as_main
-  File "<frozen runpy>", line 88, in _run_code
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 151, in <module>
-    main()
-    ~~~~^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 147, in main
-    common.main(process_treebank, common.ModelType.DEPPARSE, add_specific_args)
-    ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/common.py", line 335, in main
-    process_treebank(treebank, model_type, paths, args)
-    ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py", line 96, in process_treebank
-    prepare_tokenizer_treebank.copy_conllu_treebank(treebank, model_type, paths, paths["DEPPARSE_DATA_DIR"])
-    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_tokenizer_treebank.py", line 75, in copy_conllu_treebank
-    process_treebank(treebank, model_type, paths, args)
-    ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_tokenizer_treebank.py", line 1499, in process_treebank
-    process_ud_treebank(treebank, udbase_dir, tokenizer_dir, short_name, short_language, args.augment)
-    ~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_tokenizer_treebank.py", line 1357, in process_ud_treebank
-    prepare_ud_dataset(treebank, udbase_dir, tokenizer_dir, short_name, short_language, "test", augment)
-    ~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_tokenizer_treebank.py", line 1332, in prepare_ud_dataset
-    input_conllu = common.find_treebank_dataset_file(treebank, udbase_dir, dataset, "conllu", fail=True)
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/common.py", line 196, in find_treebank_dataset_file
-    raise FileNotFoundError("Could not find any treebank files which matched {}\nIf you have the data elsewhere, you can change the base directory for the search by changing the {} environment variable".format(filename, env_var))
-FileNotFoundError: Could not find any treebank files which matched /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/*-ud-test.conllu
-If you have the data elsewhere, you can change the base directory for the search by changing the UDBASE environment variable
-Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /tmp/tmp468rp5x_/sv_diachronic.train.gold.conllu
-Swapped 'w1, w2' for 'w1 ,w2' 93 times
-Added 86 new sentences with asdf, zzzz -> asdf,zzzz
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /tmp/tmp468rp5x_/sv_diachronic.dev.gold.conllu

logs/log_conll17.pt_sv_20260219_105228.txt

@@ -1,315 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_20260219_105228.txt ===
-Language codes: sv
-Using pretrained model: conll17.pt
-
-Running: python prepare-train-val-SILVER.py sv
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /home/urdatorn/git/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Loading DigPhil MACHINE for SILVER (minus gold overlap)…
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /home/urdatorn/git/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
-Adding 74 remaining gold sentences to TRAIN
-Cleaning TRAIN...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Cleaning SILVER...
-[REMOVED] sent_id=11_12_13  ERRORS=['Multiple roots found: [5, 17, 25]']
-[REMOVED] sent_id=65_split2_66_split2  ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
-[REMOVED] sent_id=50_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
-[REMOVED] sent_id=53_54  ERRORS=['Multiple roots found: [27, 91]']
-[REMOVED] sent_id=55_56_57  ERRORS=['Multiple roots found: [2, 4, 13]']
-[REMOVED] sent_id=909  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=912  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=396  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=416  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=589  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=17_split1  ERRORS=['Multiple roots found: [2, 14, 17]']
-[REMOVED] sent_id=17_split2  ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
-[REMOVED] sent_id=19_split1  ERRORS=['Multiple roots found: [3, 31]']
-[REMOVED] sent_id=87_88  ERRORS=['Multiple roots found: [3, 6]']
-[REMOVED] sent_id=25  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=136  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=208  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=230  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=245  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=276  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=320  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=366  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=519  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=569  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=6  ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
-[REMOVED] sent_id=7_8  ERRORS=['Multiple roots found: [5, 10]']
-[REMOVED] sent_id=124_split2  ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
-[REMOVED] sent_id=30_31  ERRORS=['Multiple roots found: [3, 18]']
-[REMOVED] sent_id=35  ERRORS=['Line 36: Invalid token ID or head']
-[REMOVED] sent_id=119  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=188  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=223  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=268  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=325  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=388  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=399  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=475  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=505  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=520  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=562  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=711  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=731  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=867  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=884  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=923  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=939  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1086  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1251  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1345  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1459  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1656  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=3_split1  ERRORS=['Multiple roots found: [4, 15, 17]']
-[REMOVED] sent_id=3_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=3_4  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=5_6  ERRORS=['Multiple roots found: [3, 24]']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [7, 20]']
-[REMOVED] sent_id=8_9  ERRORS=['Multiple roots found: [24, 57]']
-[REMOVED] sent_id=12_13  ERRORS=['Multiple roots found: [11, 16]']
-Cleaning DEV...
-Writing TRAIN → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (13867 valid sentences)
-Writing SILVER → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train-silver.conllu (46432 valid sentences)
-Writing DEV → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (35 valid sentences)
-Writing TEST → /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (copy of dev, required by Stanza)
-Done.
-Sourcing scripts/config.sh
-Running stanza dataset preparation…
-2026-02-19 10:52:31 INFO: Datasets program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt --gold
-Augmented 60 quotes: Counter({'《》': 9, '»«': 8, '″″': 8, '““': 7, '„”': 6, '""': 6, '„“': 6, '«»': 4, '””': 4, '「」': 2})
-Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /tmp/tmpwrb749kh/sv_diachronic.train.gold.conllu
-Swapped 'w1, w2' for 'w1 ,w2' 98 times
-Added 86 new sentences with asdf, zzzz -> asdf,zzzz
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /tmp/tmpwrb749kh/sv_diachronic.dev.gold.conllu
-Reading from /home/urdatorn/git/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /tmp/tmpwrb749kh/sv_diachronic.test.gold.conllu
-Copying from /tmp/tmpwrb749kh/sv_diachronic.train.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-Copying from /tmp/tmpwrb749kh/sv_diachronic.dev.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
-Copying from /tmp/tmpwrb749kh/sv_diachronic.test.gold.conllu to /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
-Running stanza dependency parser training…
-2026-02-19 10:52:34 INFO: Training program called with:
-/home/urdatorn/git/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --batch_size 32 --dropout 0.33 --use_bert --bert_model vesteinn/ScandiBERT --silver_file ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu
-2026-02-19 10:52:34 DEBUG: UD_Swedish-diachronic: sv_diachronic
-2026-02-19 10:52:34 INFO: Using model /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2026-02-19 10:52:34 INFO: Using model /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2026-02-19 10:52:34 INFO: Using default pretrain for language sv, found in /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt  To use a different pretrain, specify --wordvec_pretrain_file
-2026-02-19 10:52:34 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_scandibert_parser.pt does not exist, training new model
-2026-02-19 10:52:34 INFO: Using model /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2026-02-19 10:52:34 INFO: Using model /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2026-02-19 10:52:34 INFO: Using default pretrain for language sv, found in /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt  To use a different pretrain, specify --wordvec_pretrain_file
-2026-02-19 10:52:34 INFO: Running train depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/home/urdatorn/stanza_resources/sv/pretrain', '--train_file', '/home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu', '--eval_file', '/home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--batch_size', '5000', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'train', '--wordvec_pretrain_file', '/home/urdatorn/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt', '--batch_size', '32', '--dropout', '0.33', '--bert_model', 'vesteinn/ScandiBERT', '--silver_file', 'ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu']
-2026-02-19 10:52:34 INFO: Running parser in train mode
-2026-02-19 10:52:35 DEBUG: Loaded pretrain from /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt
-2026-02-19 10:52:35 INFO: Using 7 as the word cutoff based on the size of the pretrain (250000)
-2026-02-19 10:52:35 INFO: Using pretrained contextualized char embedding
-2026-02-19 10:52:35 INFO: ARGS USED AT TRAINING TIME:
-augment_nopunct: None
-batch_size: 32
-bert_finetune: False
-bert_finetune_layers: None
-bert_hidden_layers: 4
-bert_learning_rate: 1.0
-bert_model: vesteinn/ScandiBERT
-bert_start_finetuning: 200
-bert_warmup_steps: 200
-bert_weight_decay: 0.0
-beta2: 0.999
-char: True
-char_emb_dim: 100
-char_hidden_dim: 400
-char_num_layers: 1
-char_rec_dropout: 0
-charlm: True
-charlm_backward_file: /home/urdatorn/stanza_resources/sv/backward_charlm/conll17.pt
-charlm_forward_file: /home/urdatorn/stanza_resources/sv/forward_charlm/conll17.pt
-charlm_save_dir: saved_models/charlm
-charlm_shorthand: sv_conll17
-checkpoint: True
-checkpoint_interval: 500
-checkpoint_save_name: None
-continue_from: None
-data_dir: data/depparse
-deep_biaff_hidden_dim: 400
-deep_biaff_output_dim: 160
-device: cpu
-distance: True
-dropout: 0.33
-eval_file: /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
-eval_interval: 100
-gold_labels: True
-hidden_dim: 400
-lang: sv
-linearization: True
-log_norms: False
-log_step: 20
-lora_alpha: 128
-lora_dropout: 0.1
-lora_modules_to_save: []
-lora_rank: 64
-lora_target_modules: ['query', 'value', 'output.dense', 'intermediate.dense']
-lr: 2.0
-max_grad_norm: 1.0
-max_steps: 50000
-max_steps_before_stop: 2000
-mode: train
-model_type: graph
-num_layers: 3
-optim: adadelta
-output_file: None
-output_latex: False
-pretrain: True
-pretrain_max_vocab: 250000
-rec_dropout: 0
-reversed: False
-sample_train: 1.0
-save_dir: saved_models/depparse
-save_name: {shorthand}_{embedding}_parser.pt
-second_batch_size: None
-second_bert_learning_rate: 0.001
-second_lr: 0.0002
-second_optim: adam
-second_optim_start_step: 10000
-second_warmup_steps: 200
-seed: 1234
-shorthand: sv_diachronic
-silver_file: ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu
-silver_weight: 0.5
-tag_emb_dim: 50
-train_file: /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-train_size: None
-transformed_dim: 125
-transition_embedding_dim: 20
-transition_hidden_dim: 20
-transition_merge_hidden_dim: 200
-transition_subtree_combination: SubtreeCombination.NONE
-transition_subtree_nonlinearity: none
-use_arc_embedding: False
-use_peft: False
-use_ufeats: True
-use_upos: True
-use_xpos: True
-wandb: False
-wandb_name: None
-weight_decay: 1e-05
-word_cutoff: 7
-word_dropout: 0.33
-word_emb_dim: 75
-wordvec_dir: /home/urdatorn/stanza_resources/sv/pretrain
-wordvec_file: None
-wordvec_pretrain_file: /home/urdatorn/stanza_resources/sv/pretrain/conll17.pt
-
-2026-02-19 10:52:35 INFO: Loading data with batch size 32...
-2026-02-19 10:52:36 INFO: Train File /home/urdatorn/git/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu, Data Size: 13953
-2026-02-19 10:52:36 INFO: Original data size: 13953
-2026-02-19 10:52:38 WARNING: sv_diachronic is not a known dataset.  Examining the data to choose which xpos vocab to use
-2026-02-19 10:52:38 INFO: Original length = 13953
-2026-02-19 10:52:38 INFO: Filtered length = 13953
-2026-02-19 10:52:40 WARNING: Chose XPOSDescription(xpos_type=<XPOSType.XPOS: 1>, sep='-') for the xpos factory for sv_diachronic
-2026-02-19 10:52:42 DEBUG: 9512 batches created.
-2026-02-19 10:52:42 INFO: Augmented data size: 14711
-2026-02-19 10:52:45 DEBUG: 10020 batches created.
-2026-02-19 10:52:45 DEBUG: 34 batches created.
-2026-02-19 10:52:45 DEBUG: 34 batches created.
-2026-02-19 10:52:45 INFO: Training parser...
-
-XLMRobertaModel LOAD REPORT from: vesteinn/ScandiBERT
-Key                             | Status     | 
---------------------------------+------------+-
-lm_head.bias                    | UNEXPECTED | 
-lm_head.decoder.bias            | UNEXPECTED | 
-lm_head.layer_norm.bias         | UNEXPECTED | 
-lm_head.dense.weight            | UNEXPECTED | 
-lm_head.layer_norm.weight       | UNEXPECTED | 
-roberta.embeddings.position_ids | UNEXPECTED | 
-lm_head.dense.bias              | UNEXPECTED | 
-pooler.dense.bias               | MISSING    | 
-pooler.dense.weight             | MISSING    | 
-
-Notes:
-- UNEXPECTED	:can be ignored when loading from different task/architecture; not ok if you expect identical arch.
-- MISSING	:those params were newly initialized because missing from the checkpoint. Consider training on your downstream task.
-Warning: You are sending unauthenticated requests to the HF Hub. Please set a HF_TOKEN to enable higher rate limits and faster downloads.
-Traceback (most recent call last):
-  File "<frozen runpy>", line 198, in _run_module_as_main
-  File "<frozen runpy>", line 88, in _run_code
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/training/run_depparse.py", line 145, in <module>
-    main()
-    ~~~~^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/training/run_depparse.py", line 142, in main
-    common.main(run_treebank, "depparse", "parser", add_depparse_args, sub_argparse=parser.build_argparse(), build_model_filename=build_model_filename, choose_charlm_method=choose_depparse_charlm)
-    ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/training/common.py", line 201, in main
-    run_treebank(mode, paths, treebank, short_name, command_args, extra_args + save_name_args)
-    ~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/utils/training/run_depparse.py", line 94, in run_treebank
-    parser.main(train_args)
-    ~~~~~~~~~~~^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/models/parser.py", line 266, in main
-    return train(args)
-  File "/home/urdatorn/git/stanza-digphil/stanza/models/parser.py", line 392, in train
-    trainer = model_type(args=args, vocab=vocab, pretrain=pretrain, device=args['device'])
-  File "/home/urdatorn/git/stanza-digphil/stanza/models/depparse/trainer.py", line 65, in __init__
-    bert_model, bert_tokenizer = load_bert(self.args['bert_model'])
-                                 ~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/models/common/foundation_cache.py", line 115, in load_bert
-    return bert_embedding.load_bert(model_name, local_files_only=local_files_only)
-           ~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/git/stanza-digphil/stanza/models/common/bert_embedding.py", line 66, in load_bert
-    bert_tokenizer = load_tokenizer(model_name, tokenizer_kwargs=tokenizer_kwargs, local_files_only=local_files_only)
-  File "/home/urdatorn/git/stanza-digphil/stanza/models/common/bert_embedding.py", line 50, in load_tokenizer
-    bert_tokenizer = AutoTokenizer.from_pretrained(model_name, **bert_args)
-  File "/home/urdatorn/.pyenv/versions/venv/lib/python3.14/site-packages/transformers/models/auto/tokenization_auto.py", line 712, in from_pretrained
-    return tokenizer_class.from_pretrained(pretrained_model_name_or_path, *inputs, **kwargs)
-           ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-  File "/home/urdatorn/.pyenv/versions/venv/lib/python3.14/site-packages/transformers/tokenization_utils_base.py", line 1712, in from_pretrained
-    return cls._from_pretrained(
-           ~~~~~~~~~~~~~~~~~~~~^
-        resolved_vocab_files,
-        ^^^^^^^^^^^^^^^^^^^^^
-    ...<9 lines>...
-        **kwargs,
-        ^^^^^^^^^
-    )
-    ^
-  File "/home/urdatorn/.pyenv/versions/venv/lib/python3.14/site-packages/transformers/tokenization_utils_base.py", line 1897, in _from_pretrained
-    init_kwargs = cls.convert_to_native_format(**init_kwargs)
-  File "/home/urdatorn/.pyenv/versions/venv/lib/python3.14/site-packages/transformers/tokenization_utils_tokenizers.py", line 127, in convert_to_native_format
-    if vocab and isinstance(vocab[0], (list, tuple)):
-                            ~~~~~^^^
-KeyError: 0

logs/log_conll17.pt_sv_diachron_20251203_223822.txt

@@ -1,731 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_diachron_20251203_223822.txt ===
-Language codes: sv diachron
-Using pretrained model: conll17.pt
-
-Running: python prepare-train-val-test.py sv diachron
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/ucxn_ud_swedish-talbanken.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Including DigPhil MACHINE in TRAIN (minus gold)…
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
-Cleaning TRAIN...
-[REMOVED] sent_id=6  ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
-[REMOVED] sent_id=7_8  ERRORS=['Multiple roots found: [5, 10]']
-[REMOVED] sent_id=30_31  ERRORS=['Multiple roots found: [3, 18]']
-[REMOVED] sent_id=35  ERRORS=['Line 36: Invalid token ID or head']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [7, 20]']
-[REMOVED] sent_id=8_9  ERRORS=['Multiple roots found: [24, 57]']
-[REMOVED] sent_id=12_13  ERRORS=['Multiple roots found: [11, 16]']
-[REMOVED] sent_id=124_split2  ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
-[REMOVED] sent_id=396  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=416  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=589  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=909  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=912  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=3_split1  ERRORS=['Multiple roots found: [4, 15, 17]']
-[REMOVED] sent_id=3_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
-[REMOVED] sent_id=3_4  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=5_6  ERRORS=['Multiple roots found: [3, 24]']
-[REMOVED] sent_id=11_12_13  ERRORS=['Multiple roots found: [5, 17, 25]']
-[REMOVED] sent_id=119  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=188  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=223  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=268  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=325  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=388  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=399  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=475  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=505  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=520  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=562  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=711  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=731  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=867  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=884  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=923  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=939  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1086  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1251  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1345  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1459  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1656  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=87_88  ERRORS=['Multiple roots found: [3, 6]']
-[REMOVED] sent_id=65_split2_66_split2  ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
-[REMOVED] sent_id=25  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=136  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=208  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=230  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=245  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=276  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=320  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=366  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=519  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=569  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=50_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
-[REMOVED] sent_id=53_54  ERRORS=['Multiple roots found: [27, 91]']
-[REMOVED] sent_id=55_56_57  ERRORS=['Multiple roots found: [2, 4, 13]']
-[REMOVED] sent_id=17_split1  ERRORS=['Multiple roots found: [2, 14, 17]']
-[REMOVED] sent_id=17_split2  ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
-[REMOVED] sent_id=19_split1  ERRORS=['Multiple roots found: [3, 31]']
-Cleaning DEV...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Cleaning TEST...
-Writing TRAIN → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (66252 valid sentences)
-Writing DEV → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (9 valid sentences)
-Writing TEST → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (99 valid sentences)
-Done.
-Sourcing scripts/config_alvis.sh
-Running stanza dataset preparation…
-2025-12-03 22:38:31 INFO: Datasets program called with:
-/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 22:38:31 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
-
-2025-12-03 22:38:31 INFO: Downloaded file to /cephyr/users/cleland/Alvis/stanza_resources/resources.json
-2025-12-03 22:38:31 DEBUG: Processing parameter "processors"...
-2025-12-03 22:38:31 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
-2025-12-03 22:38:31 INFO: Downloading these customized packages for language: sv (Swedish)...
-=======================
-| Processor | Package |
------------------------
-=======================
-
-2025-12-03 22:38:31 INFO: Finished downloading models and saved to /cephyr/users/cleland/Alvis/stanza_resources
-2025-12-03 22:38:31 INFO: Using tagger model in /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt for sv_diachronic
-2025-12-03 22:38:31 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2025-12-03 22:38:31 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-Augmented 192 quotes: Counter({'„”': 28, '""': 27, '「」': 20, '″″': 20, '»«': 18, '«»': 18, '《》': 17, '””': 17, '„“': 15, '““': 12})
-2025-12-03 22:38:34 INFO: Running tagger to retag /local/tmp.5441282/tmpi3ipyceb/sv_diachronic.train.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5441282/tmpi3ipyceb/sv_diachronic.train.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu']
-2025-12-03 22:38:34 INFO: Running tagger in predict mode
-2025-12-03 22:38:34 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2025-12-03 22:38:36 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 22:38:36 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 22:38:36 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 22:38:36 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 22:38:37 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 22:38:39 INFO: Loading data with batch size 250...
-2025-12-03 22:39:18 INFO: Start evaluation...
-2025-12-03 22:42:53 INFO: UPOS	XPOS	UFeats	AllTags
-2025-12-03 22:42:53 INFO: 99.37	88.39	98.13	87.63
-2025-12-03 22:42:53 INFO: POS Tagger score: sv_diachronic 87.63
-2025-12-03 22:42:53 INFO: Running tagger to retag /local/tmp.5441282/tmpi3ipyceb/sv_diachronic.dev.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5441282/tmpi3ipyceb/sv_diachronic.dev.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu']
-2025-12-03 22:42:53 INFO: Running tagger in predict mode
-2025-12-03 22:42:53 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2025-12-03 22:42:55 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 22:42:55 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 22:42:55 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 22:42:55 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 22:42:55 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 22:42:55 INFO: Loading data with batch size 250...
-2025-12-03 22:42:55 INFO: Start evaluation...
-2025-12-03 22:42:56 INFO: UPOS	XPOS	UFeats	AllTags
-2025-12-03 22:42:56 INFO: 93.32	90.84	93.32	85.64
-2025-12-03 22:42:56 INFO: POS Tagger score: sv_diachronic 85.64
-2025-12-03 22:42:56 INFO: Running tagger to retag /local/tmp.5441282/tmpi3ipyceb/sv_diachronic.test.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5441282/tmpi3ipyceb/sv_diachronic.test.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu']
-2025-12-03 22:42:56 INFO: Running tagger in predict mode
-2025-12-03 22:42:56 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2025-12-03 22:42:57 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 22:42:57 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 22:42:57 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 22:42:57 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 22:42:58 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 22:42:58 INFO: Loading data with batch size 250...
-2025-12-03 22:42:58 INFO: Start evaluation...
-2025-12-03 22:42:58 INFO: UPOS	XPOS	UFeats	AllTags
-2025-12-03 22:42:58 INFO: 93.14	96.78	95.32	90.28
-2025-12-03 22:42:58 INFO: POS Tagger score: sv_diachronic 90.28
-Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /local/tmp.5441282/tmpi3ipyceb/sv_diachronic.train.gold.conllu
-Swapped 'w1, w2' for 'w1 ,w2' 132 times
-Added 506 new sentences with asdf, zzzz -> asdf,zzzz
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /local/tmp.5441282/tmpi3ipyceb/sv_diachronic.dev.gold.conllu
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /local/tmp.5441282/tmpi3ipyceb/sv_diachronic.test.gold.conllu
-Running stanza dependency parser training…
-2025-12-03 22:43:12 INFO: Training program called with:
-/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt --batch_size 32 --dropout 0.33
-2025-12-03 22:43:12 DEBUG: UD_Swedish-diachronic: sv_diachronic
-2025-12-03 22:43:12 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2025-12-03 22:43:12 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2025-12-03 22:43:12 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_charlm_parser.pt does not exist, training new model
-2025-12-03 22:43:12 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2025-12-03 22:43:12 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2025-12-03 22:43:12 INFO: Running train depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--train_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--batch_size', '5000', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'train', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
-2025-12-03 22:43:12 INFO: Running parser in train mode
-2025-12-03 22:43:12 INFO: Using pretrained contextualized char embedding
-2025-12-03 22:43:12 INFO: Loading data with batch size 32...
-2025-12-03 22:43:20 INFO: Train File /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu, Data Size: 66758
-2025-12-03 22:43:20 INFO: Original data size: 66758
-2025-12-03 22:43:21 INFO: Augmented data size: 66875
-2025-12-03 22:43:38 WARNING: sv_diachronic is not a known dataset.  Examining the data to choose which xpos vocab to use
-2025-12-03 22:43:38 INFO: Original length = 66875
-2025-12-03 22:43:38 INFO: Filtered length = 66875
-2025-12-03 22:43:55 WARNING: Chose XPOSDescription(xpos_type=<XPOSType.XPOS: 1>, sep='|') for the xpos factory for sv_diachronic
-2025-12-03 22:44:01 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 22:44:15 DEBUG: 38839 batches created.
-2025-12-03 22:44:15 DEBUG: 9 batches created.
-2025-12-03 22:44:15 INFO: Training parser...
-2025-12-03 22:44:15 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 22:44:15 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 22:44:15 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 22:44:16 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 22:44:24 INFO: Finished STEP 20/50000, loss = 6.296603 (0.201 sec/batch), lr: 0.003000
-2025-12-03 22:44:27 INFO: Finished STEP 40/50000, loss = 5.081994 (0.181 sec/batch), lr: 0.003000
-2025-12-03 22:44:31 INFO: Finished STEP 60/50000, loss = 4.866196 (0.144 sec/batch), lr: 0.003000
-2025-12-03 22:44:34 INFO: Finished STEP 80/50000, loss = 4.271401 (0.138 sec/batch), lr: 0.003000
-2025-12-03 22:44:36 INFO: Finished STEP 100/50000, loss = 4.687382 (0.143 sec/batch), lr: 0.003000
-2025-12-03 22:44:36 INFO: Evaluating on dev set...
-2025-12-03 22:44:37 INFO: LAS	MLAS	BLEX
-2025-12-03 22:44:37 INFO: 43.56	33.47	35.15
-2025-12-03 22:44:37 INFO: step 100: train_loss = 7.396650, dev_score = 0.4356
-2025-12-03 22:44:37 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 22:44:37 INFO: new best model saved.
-2025-12-03 22:44:38 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:44:38 INFO: new model checkpoint saved.
-2025-12-03 22:44:41 INFO: Finished STEP 120/50000, loss = 3.784610 (0.125 sec/batch), lr: 0.003000
-2025-12-03 22:44:43 INFO: Finished STEP 140/50000, loss = 4.674234 (0.131 sec/batch), lr: 0.003000
-2025-12-03 22:44:46 INFO: Finished STEP 160/50000, loss = 4.012815 (0.137 sec/batch), lr: 0.003000
-2025-12-03 22:44:48 INFO: Finished STEP 180/50000, loss = 4.966728 (0.122 sec/batch), lr: 0.003000
-2025-12-03 22:44:51 INFO: Finished STEP 200/50000, loss = 4.140534 (0.121 sec/batch), lr: 0.003000
-2025-12-03 22:44:51 INFO: Evaluating on dev set...
-2025-12-03 22:44:51 INFO: LAS	MLAS	BLEX
-2025-12-03 22:44:51 INFO: 51.98	38.30	43.83
-2025-12-03 22:44:51 INFO: step 200: train_loss = 4.683159, dev_score = 0.5198
-2025-12-03 22:44:51 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 22:44:51 INFO: new best model saved.
-2025-12-03 22:44:52 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:44:52 INFO: new model checkpoint saved.
-2025-12-03 22:44:55 INFO: Finished STEP 220/50000, loss = 3.998342 (0.126 sec/batch), lr: 0.003000
-2025-12-03 22:44:57 INFO: Finished STEP 240/50000, loss = 5.531576 (0.124 sec/batch), lr: 0.003000
-2025-12-03 22:44:59 INFO: Finished STEP 260/50000, loss = 4.806288 (0.116 sec/batch), lr: 0.003000
-2025-12-03 22:45:02 INFO: Finished STEP 280/50000, loss = 4.965835 (0.111 sec/batch), lr: 0.003000
-2025-12-03 22:45:04 INFO: Finished STEP 300/50000, loss = 3.812420 (0.115 sec/batch), lr: 0.003000
-2025-12-03 22:45:04 INFO: Evaluating on dev set...
-2025-12-03 22:45:04 INFO: LAS	MLAS	BLEX
-2025-12-03 22:45:04 INFO: 55.45	44.40	49.47
-2025-12-03 22:45:04 INFO: step 300: train_loss = 4.276576, dev_score = 0.5545
-2025-12-03 22:45:05 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 22:45:05 INFO: new best model saved.
-2025-12-03 22:45:05 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:45:05 INFO: new model checkpoint saved.
-2025-12-03 22:45:07 INFO: Finished STEP 320/50000, loss = 3.429461 (0.111 sec/batch), lr: 0.003000
-2025-12-03 22:45:10 INFO: Finished STEP 340/50000, loss = 3.642277 (0.118 sec/batch), lr: 0.003000
-2025-12-03 22:45:12 INFO: Finished STEP 360/50000, loss = 3.760803 (0.116 sec/batch), lr: 0.003000
-2025-12-03 22:45:14 INFO: Finished STEP 380/50000, loss = 4.786219 (0.105 sec/batch), lr: 0.003000
-2025-12-03 22:45:16 INFO: Finished STEP 400/50000, loss = 2.972147 (0.108 sec/batch), lr: 0.003000
-2025-12-03 22:45:16 INFO: Evaluating on dev set...
-2025-12-03 22:45:17 INFO: LAS	MLAS	BLEX
-2025-12-03 22:45:17 INFO: 57.43	46.61	50.85
-2025-12-03 22:45:17 INFO: step 400: train_loss = 3.942294, dev_score = 0.5743
-2025-12-03 22:45:17 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 22:45:17 INFO: new best model saved.
-2025-12-03 22:45:18 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:45:18 INFO: new model checkpoint saved.
-2025-12-03 22:45:20 INFO: Finished STEP 420/50000, loss = 4.565236 (0.105 sec/batch), lr: 0.003000
-2025-12-03 22:45:22 INFO: Finished STEP 440/50000, loss = 4.531408 (0.098 sec/batch), lr: 0.003000
-2025-12-03 22:45:24 INFO: Finished STEP 460/50000, loss = 5.082150 (0.090 sec/batch), lr: 0.003000
-2025-12-03 22:45:26 INFO: Finished STEP 480/50000, loss = 2.923217 (0.110 sec/batch), lr: 0.003000
-2025-12-03 22:45:28 INFO: Finished STEP 500/50000, loss = 3.408716 (0.100 sec/batch), lr: 0.003000
-2025-12-03 22:45:28 INFO: Evaluating on dev set...
-2025-12-03 22:45:28 INFO: LAS	MLAS	BLEX
-2025-12-03 22:45:28 INFO: 60.89	51.48	55.27
-2025-12-03 22:45:28 INFO: step 500: train_loss = 4.009553, dev_score = 0.6089
-2025-12-03 22:45:29 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 22:45:29 INFO: new best model saved.
-2025-12-03 22:45:29 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:45:29 INFO: new model checkpoint saved.
-2025-12-03 22:45:31 INFO: Finished STEP 520/50000, loss = 5.306584 (0.101 sec/batch), lr: 0.003000
-2025-12-03 22:45:33 INFO: Finished STEP 540/50000, loss = 3.657916 (0.100 sec/batch), lr: 0.003000
-2025-12-03 22:45:35 INFO: Finished STEP 560/50000, loss = 4.001961 (0.086 sec/batch), lr: 0.003000
-2025-12-03 22:45:37 INFO: Finished STEP 580/50000, loss = 4.052956 (0.085 sec/batch), lr: 0.003000
-2025-12-03 22:45:39 INFO: Finished STEP 600/50000, loss = 3.852856 (0.099 sec/batch), lr: 0.003000
-2025-12-03 22:45:39 INFO: Evaluating on dev set...
-2025-12-03 22:45:40 INFO: LAS	MLAS	BLEX
-2025-12-03 22:45:40 INFO: 59.90	50.00	55.51
-2025-12-03 22:45:40 INFO: step 600: train_loss = 3.724004, dev_score = 0.5990
-2025-12-03 22:45:40 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:45:40 INFO: new model checkpoint saved.
-2025-12-03 22:45:42 INFO: Finished STEP 620/50000, loss = 3.484413 (0.101 sec/batch), lr: 0.003000
-2025-12-03 22:45:44 INFO: Finished STEP 640/50000, loss = 3.220206 (0.102 sec/batch), lr: 0.003000
-2025-12-03 22:45:46 INFO: Finished STEP 660/50000, loss = 3.800846 (0.094 sec/batch), lr: 0.003000
-2025-12-03 22:45:48 INFO: Finished STEP 680/50000, loss = 3.769488 (0.093 sec/batch), lr: 0.003000
-2025-12-03 22:45:50 INFO: Finished STEP 700/50000, loss = 3.285978 (0.097 sec/batch), lr: 0.003000
-2025-12-03 22:45:50 INFO: Evaluating on dev set...
-2025-12-03 22:45:50 INFO: LAS	MLAS	BLEX
-2025-12-03 22:45:50 INFO: 57.92	46.93	53.70
-2025-12-03 22:45:50 INFO: step 700: train_loss = 3.829402, dev_score = 0.5792
-2025-12-03 22:45:51 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:45:51 INFO: new model checkpoint saved.
-2025-12-03 22:45:53 INFO: Finished STEP 720/50000, loss = 4.905636 (0.100 sec/batch), lr: 0.003000
-2025-12-03 22:45:55 INFO: Finished STEP 740/50000, loss = 3.337056 (0.102 sec/batch), lr: 0.003000
-2025-12-03 22:45:57 INFO: Finished STEP 760/50000, loss = 5.074053 (0.092 sec/batch), lr: 0.003000
-2025-12-03 22:45:59 INFO: Finished STEP 780/50000, loss = 4.024767 (0.089 sec/batch), lr: 0.003000
-2025-12-03 22:46:00 INFO: Finished STEP 800/50000, loss = 4.552472 (0.091 sec/batch), lr: 0.003000
-2025-12-03 22:46:00 INFO: Evaluating on dev set...
-2025-12-03 22:46:01 INFO: LAS	MLAS	BLEX
-2025-12-03 22:46:01 INFO: 61.39	49.05	53.70
-2025-12-03 22:46:01 INFO: step 800: train_loss = 3.612613, dev_score = 0.6139
-2025-12-03 22:46:01 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 22:46:01 INFO: new best model saved.
-2025-12-03 22:46:02 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:46:02 INFO: new model checkpoint saved.
-2025-12-03 22:46:04 INFO: Finished STEP 820/50000, loss = 6.260066 (0.090 sec/batch), lr: 0.003000
-2025-12-03 22:46:05 INFO: Finished STEP 840/50000, loss = 4.288107 (0.092 sec/batch), lr: 0.003000
-2025-12-03 22:46:07 INFO: Finished STEP 860/50000, loss = 3.849105 (0.097 sec/batch), lr: 0.003000
-2025-12-03 22:46:09 INFO: Finished STEP 880/50000, loss = 2.492743 (0.088 sec/batch), lr: 0.003000
-2025-12-03 22:46:11 INFO: Finished STEP 900/50000, loss = 4.095746 (0.090 sec/batch), lr: 0.003000
-2025-12-03 22:46:11 INFO: Evaluating on dev set...
-2025-12-03 22:46:11 INFO: LAS	MLAS	BLEX
-2025-12-03 22:46:11 INFO: 60.40	51.68	55.04
-2025-12-03 22:46:11 INFO: step 900: train_loss = 3.946219, dev_score = 0.6040
-2025-12-03 22:46:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:46:12 INFO: new model checkpoint saved.
-2025-12-03 22:46:14 INFO: Finished STEP 920/50000, loss = 2.770604 (0.084 sec/batch), lr: 0.003000
-2025-12-03 22:46:16 INFO: Finished STEP 940/50000, loss = 3.612128 (0.100 sec/batch), lr: 0.003000
-2025-12-03 22:46:17 INFO: Finished STEP 960/50000, loss = 4.249920 (0.089 sec/batch), lr: 0.003000
-2025-12-03 22:46:19 INFO: Finished STEP 980/50000, loss = 2.418294 (0.091 sec/batch), lr: 0.003000
-2025-12-03 22:46:21 INFO: Finished STEP 1000/50000, loss = 4.842584 (0.094 sec/batch), lr: 0.003000
-2025-12-03 22:46:21 INFO: Evaluating on dev set...
-2025-12-03 22:46:21 INFO: LAS	MLAS	BLEX
-2025-12-03 22:46:21 INFO: 61.39	53.70	56.66
-2025-12-03 22:46:21 INFO: step 1000: train_loss = 3.696227, dev_score = 0.6139
-2025-12-03 22:46:22 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 22:46:22 INFO: new best model saved.
-2025-12-03 22:46:22 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:46:22 INFO: new model checkpoint saved.
-2025-12-03 22:46:24 INFO: Finished STEP 1020/50000, loss = 3.346077 (0.088 sec/batch), lr: 0.003000
-2025-12-03 22:46:26 INFO: Finished STEP 1040/50000, loss = 3.296835 (0.087 sec/batch), lr: 0.003000
-2025-12-03 22:46:28 INFO: Finished STEP 1060/50000, loss = 2.515590 (0.088 sec/batch), lr: 0.003000
-2025-12-03 22:46:29 INFO: Finished STEP 1080/50000, loss = 3.913376 (0.075 sec/batch), lr: 0.003000
-2025-12-03 22:46:31 INFO: Finished STEP 1100/50000, loss = 5.241524 (0.077 sec/batch), lr: 0.003000
-2025-12-03 22:46:31 INFO: Evaluating on dev set...
-2025-12-03 22:46:31 INFO: LAS	MLAS	BLEX
-2025-12-03 22:46:31 INFO: 59.65	49.47	53.28
-2025-12-03 22:46:31 INFO: step 1100: train_loss = 3.721023, dev_score = 0.5965
-2025-12-03 22:46:32 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:46:32 INFO: new model checkpoint saved.
-2025-12-03 22:46:34 INFO: Finished STEP 1120/50000, loss = 2.731287 (0.086 sec/batch), lr: 0.003000
-2025-12-03 22:46:36 INFO: Finished STEP 1140/50000, loss = 3.536034 (0.080 sec/batch), lr: 0.003000
-2025-12-03 22:46:37 INFO: Finished STEP 1160/50000, loss = 3.398331 (0.079 sec/batch), lr: 0.003000
-2025-12-03 22:46:39 INFO: Finished STEP 1180/50000, loss = 5.028436 (0.078 sec/batch), lr: 0.003000
-2025-12-03 22:46:41 INFO: Finished STEP 1200/50000, loss = 3.061586 (0.077 sec/batch), lr: 0.003000
-2025-12-03 22:46:41 INFO: Evaluating on dev set...
-2025-12-03 22:46:41 INFO: LAS	MLAS	BLEX
-2025-12-03 22:46:41 INFO: 61.14	53.28	56.66
-2025-12-03 22:46:41 INFO: step 1200: train_loss = 3.522418, dev_score = 0.6114
-2025-12-03 22:46:42 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:46:42 INFO: new model checkpoint saved.
-2025-12-03 22:46:43 INFO: Finished STEP 1220/50000, loss = 2.513215 (0.089 sec/batch), lr: 0.003000
-2025-12-03 22:46:45 INFO: Finished STEP 1240/50000, loss = 4.864305 (0.085 sec/batch), lr: 0.003000
-2025-12-03 22:46:47 INFO: Finished STEP 1260/50000, loss = 7.466803 (0.082 sec/batch), lr: 0.003000
-2025-12-03 22:46:49 INFO: Finished STEP 1280/50000, loss = 3.586653 (0.086 sec/batch), lr: 0.003000
-2025-12-03 22:46:50 INFO: Finished STEP 1300/50000, loss = 3.791464 (0.078 sec/batch), lr: 0.003000
-2025-12-03 22:46:50 INFO: Evaluating on dev set...
-2025-12-03 22:46:51 INFO: LAS	MLAS	BLEX
-2025-12-03 22:46:51 INFO: 61.63	52.01	55.81
-2025-12-03 22:46:51 INFO: step 1300: train_loss = 3.743181, dev_score = 0.6163
-2025-12-03 22:46:51 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 22:46:51 INFO: new best model saved.
-2025-12-03 22:46:51 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:46:51 INFO: new model checkpoint saved.
-2025-12-03 22:46:53 INFO: Finished STEP 1320/50000, loss = 2.740164 (0.083 sec/batch), lr: 0.003000
-2025-12-03 22:46:55 INFO: Finished STEP 1340/50000, loss = 4.084477 (0.088 sec/batch), lr: 0.003000
-2025-12-03 22:46:56 INFO: Finished STEP 1360/50000, loss = 4.818987 (0.070 sec/batch), lr: 0.003000
-2025-12-03 22:46:58 INFO: Finished STEP 1380/50000, loss = 2.790164 (0.076 sec/batch), lr: 0.003000
-2025-12-03 22:47:00 INFO: Finished STEP 1400/50000, loss = 4.602959 (0.081 sec/batch), lr: 0.003000
-2025-12-03 22:47:00 INFO: Evaluating on dev set...
-2025-12-03 22:47:00 INFO: LAS	MLAS	BLEX
-2025-12-03 22:47:00 INFO: 62.13	55.08	58.05
-2025-12-03 22:47:00 INFO: step 1400: train_loss = 3.724900, dev_score = 0.6213
-2025-12-03 22:47:00 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 22:47:00 INFO: new best model saved.
-2025-12-03 22:47:01 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:47:01 INFO: new model checkpoint saved.
-2025-12-03 22:47:03 INFO: Finished STEP 1420/50000, loss = 4.015442 (0.079 sec/batch), lr: 0.003000
-2025-12-03 22:47:04 INFO: Finished STEP 1440/50000, loss = 3.268615 (0.082 sec/batch), lr: 0.003000
-2025-12-03 22:47:06 INFO: Finished STEP 1460/50000, loss = 3.735999 (0.080 sec/batch), lr: 0.003000
-2025-12-03 22:47:07 INFO: Finished STEP 1480/50000, loss = 3.804015 (0.073 sec/batch), lr: 0.003000
-2025-12-03 22:47:09 INFO: Finished STEP 1500/50000, loss = 4.517982 (0.087 sec/batch), lr: 0.003000
-2025-12-03 22:47:09 INFO: Evaluating on dev set...
-2025-12-03 22:47:10 INFO: LAS	MLAS	BLEX
-2025-12-03 22:47:10 INFO: 61.63	52.54	56.36
-2025-12-03 22:47:10 INFO: step 1500: train_loss = 3.783271, dev_score = 0.6163
-2025-12-03 22:47:10 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:47:10 INFO: new model checkpoint saved.
-2025-12-03 22:47:12 INFO: Finished STEP 1520/50000, loss = 3.094217 (0.085 sec/batch), lr: 0.003000
-2025-12-03 22:47:13 INFO: Finished STEP 1540/50000, loss = 3.673062 (0.073 sec/batch), lr: 0.003000
-2025-12-03 22:47:15 INFO: Finished STEP 1560/50000, loss = 3.584330 (0.078 sec/batch), lr: 0.003000
-2025-12-03 22:47:16 INFO: Finished STEP 1580/50000, loss = 3.714322 (0.078 sec/batch), lr: 0.003000
-2025-12-03 22:47:18 INFO: Finished STEP 1600/50000, loss = 3.680001 (0.070 sec/batch), lr: 0.003000
-2025-12-03 22:47:18 INFO: Evaluating on dev set...
-2025-12-03 22:47:18 INFO: LAS	MLAS	BLEX
-2025-12-03 22:47:18 INFO: 60.89	49.47	54.55
-2025-12-03 22:47:18 INFO: step 1600: train_loss = 3.629141, dev_score = 0.6089
-2025-12-03 22:47:19 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:47:19 INFO: new model checkpoint saved.
-2025-12-03 22:47:21 INFO: Finished STEP 1620/50000, loss = 4.405625 (0.076 sec/batch), lr: 0.003000
-2025-12-03 22:47:22 INFO: Finished STEP 1640/50000, loss = 6.583941 (0.084 sec/batch), lr: 0.003000
-2025-12-03 22:47:24 INFO: Finished STEP 1660/50000, loss = 2.624713 (0.082 sec/batch), lr: 0.003000
-2025-12-03 22:47:25 INFO: Finished STEP 1680/50000, loss = 3.746646 (0.081 sec/batch), lr: 0.003000
-2025-12-03 22:47:27 INFO: Finished STEP 1700/50000, loss = 3.647429 (0.078 sec/batch), lr: 0.003000
-2025-12-03 22:47:27 INFO: Evaluating on dev set...
-2025-12-03 22:47:27 INFO: LAS	MLAS	BLEX
-2025-12-03 22:47:27 INFO: 62.38	53.59	56.96
-2025-12-03 22:47:27 INFO: step 1700: train_loss = 3.711031, dev_score = 0.6238
-2025-12-03 22:47:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 22:47:28 INFO: new best model saved.
-2025-12-03 22:47:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:47:28 INFO: new model checkpoint saved.
-2025-12-03 22:47:30 INFO: Finished STEP 1720/50000, loss = 3.788727 (0.082 sec/batch), lr: 0.003000
-2025-12-03 22:47:31 INFO: Finished STEP 1740/50000, loss = 4.010241 (0.075 sec/batch), lr: 0.003000
-2025-12-03 22:47:33 INFO: Finished STEP 1760/50000, loss = 3.686716 (0.081 sec/batch), lr: 0.003000
-2025-12-03 22:47:35 INFO: Finished STEP 1780/50000, loss = 2.884626 (0.079 sec/batch), lr: 0.003000
-2025-12-03 22:47:36 INFO: Finished STEP 1800/50000, loss = 3.530944 (0.073 sec/batch), lr: 0.003000
-2025-12-03 22:47:36 INFO: Evaluating on dev set...
-2025-12-03 22:47:37 INFO: LAS	MLAS	BLEX
-2025-12-03 22:47:37 INFO: 63.37	53.81	58.47
-2025-12-03 22:47:37 INFO: step 1800: train_loss = 3.627865, dev_score = 0.6337
-2025-12-03 22:47:37 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 22:47:37 INFO: new best model saved.
-2025-12-03 22:47:37 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:47:37 INFO: new model checkpoint saved.
-2025-12-03 22:47:39 INFO: Finished STEP 1820/50000, loss = 3.387048 (0.078 sec/batch), lr: 0.003000
-2025-12-03 22:47:40 INFO: Finished STEP 1840/50000, loss = 3.075635 (0.081 sec/batch), lr: 0.003000
-2025-12-03 22:47:42 INFO: Finished STEP 1860/50000, loss = 3.544627 (0.074 sec/batch), lr: 0.003000
-2025-12-03 22:47:44 INFO: Finished STEP 1880/50000, loss = 2.697122 (0.073 sec/batch), lr: 0.003000
-2025-12-03 22:47:45 INFO: Finished STEP 1900/50000, loss = 4.582170 (0.074 sec/batch), lr: 0.003000
-2025-12-03 22:47:45 INFO: Evaluating on dev set...
-2025-12-03 22:47:46 INFO: LAS	MLAS	BLEX
-2025-12-03 22:47:46 INFO: 63.12	55.81	57.93
-2025-12-03 22:47:46 INFO: step 1900: train_loss = 3.833838, dev_score = 0.6312
-2025-12-03 22:47:46 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:47:46 INFO: new model checkpoint saved.
-2025-12-03 22:47:48 INFO: Finished STEP 1920/50000, loss = 4.797754 (0.071 sec/batch), lr: 0.003000
-2025-12-03 22:47:49 INFO: Finished STEP 1940/50000, loss = 3.999227 (0.076 sec/batch), lr: 0.003000
-2025-12-03 22:47:51 INFO: Finished STEP 1960/50000, loss = 2.863396 (0.081 sec/batch), lr: 0.003000
-2025-12-03 22:47:52 INFO: Finished STEP 1980/50000, loss = 3.603798 (0.074 sec/batch), lr: 0.003000
-2025-12-03 22:47:54 INFO: Finished STEP 2000/50000, loss = 2.377973 (0.075 sec/batch), lr: 0.003000
-2025-12-03 22:47:54 INFO: Evaluating on dev set...
-2025-12-03 22:47:54 INFO: LAS	MLAS	BLEX
-2025-12-03 22:47:54 INFO: 58.91	50.63	54.85
-2025-12-03 22:47:54 INFO: step 2000: train_loss = 3.578752, dev_score = 0.5891
-2025-12-03 22:47:55 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:47:55 INFO: new model checkpoint saved.
-2025-12-03 22:47:56 INFO: Finished STEP 2020/50000, loss = 3.455811 (0.069 sec/batch), lr: 0.003000
-2025-12-03 22:47:58 INFO: Finished STEP 2040/50000, loss = 4.113457 (0.067 sec/batch), lr: 0.003000
-2025-12-03 22:47:59 INFO: Finished STEP 2060/50000, loss = 3.953318 (0.070 sec/batch), lr: 0.003000
-2025-12-03 22:48:01 INFO: Finished STEP 2080/50000, loss = 2.824056 (0.075 sec/batch), lr: 0.003000
-2025-12-03 22:48:02 INFO: Finished STEP 2100/50000, loss = 4.579782 (0.079 sec/batch), lr: 0.003000
-2025-12-03 22:48:02 INFO: Evaluating on dev set...
-2025-12-03 22:48:03 INFO: LAS	MLAS	BLEX
-2025-12-03 22:48:03 INFO: 61.63	52.85	57.51
-2025-12-03 22:48:03 INFO: step 2100: train_loss = 3.729740, dev_score = 0.6163
-2025-12-03 22:48:03 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:48:03 INFO: new model checkpoint saved.
-2025-12-03 22:48:05 INFO: Finished STEP 2120/50000, loss = 3.769734 (0.077 sec/batch), lr: 0.003000
-2025-12-03 22:48:06 INFO: Finished STEP 2140/50000, loss = 3.904819 (0.071 sec/batch), lr: 0.003000
-2025-12-03 22:48:08 INFO: Finished STEP 2160/50000, loss = 3.850474 (0.068 sec/batch), lr: 0.003000
-2025-12-03 22:48:09 INFO: Finished STEP 2180/50000, loss = 3.001653 (0.072 sec/batch), lr: 0.003000
-2025-12-03 22:48:11 INFO: Finished STEP 2200/50000, loss = 1.957051 (0.074 sec/batch), lr: 0.003000
-2025-12-03 22:48:11 INFO: Evaluating on dev set...
-2025-12-03 22:48:11 INFO: LAS	MLAS	BLEX
-2025-12-03 22:48:11 INFO: 60.64	51.16	55.39
-2025-12-03 22:48:11 INFO: step 2200: train_loss = 3.531071, dev_score = 0.6064
-2025-12-03 22:48:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:48:12 INFO: new model checkpoint saved.
-2025-12-03 22:48:13 INFO: Finished STEP 2220/50000, loss = 3.429658 (0.077 sec/batch), lr: 0.003000
-2025-12-03 22:48:15 INFO: Finished STEP 2240/50000, loss = 3.791333 (0.073 sec/batch), lr: 0.003000
-2025-12-03 22:48:16 INFO: Finished STEP 2260/50000, loss = 2.263699 (0.072 sec/batch), lr: 0.003000
-2025-12-03 22:48:18 INFO: Finished STEP 2280/50000, loss = 3.650295 (0.067 sec/batch), lr: 0.003000
-2025-12-03 22:48:19 INFO: Finished STEP 2300/50000, loss = 4.082314 (0.071 sec/batch), lr: 0.003000
-2025-12-03 22:48:19 INFO: Evaluating on dev set...
-2025-12-03 22:48:20 INFO: LAS	MLAS	BLEX
-2025-12-03 22:48:20 INFO: 62.13	54.55	57.93
-2025-12-03 22:48:20 INFO: step 2300: train_loss = 3.777273, dev_score = 0.6213
-2025-12-03 22:48:20 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:48:20 INFO: new model checkpoint saved.
-2025-12-03 22:48:22 INFO: Finished STEP 2320/50000, loss = 3.704285 (0.071 sec/batch), lr: 0.003000
-2025-12-03 22:48:23 INFO: Finished STEP 2340/50000, loss = 2.753342 (0.069 sec/batch), lr: 0.003000
-2025-12-03 22:48:25 INFO: Finished STEP 2360/50000, loss = 3.819938 (0.072 sec/batch), lr: 0.003000
-2025-12-03 22:48:26 INFO: Finished STEP 2380/50000, loss = 4.015243 (0.075 sec/batch), lr: 0.003000
-2025-12-03 22:48:27 INFO: Finished STEP 2400/50000, loss = 4.291789 (0.077 sec/batch), lr: 0.003000
-2025-12-03 22:48:27 INFO: Evaluating on dev set...
-2025-12-03 22:48:28 INFO: LAS	MLAS	BLEX
-2025-12-03 22:48:28 INFO: 60.64	53.28	57.08
-2025-12-03 22:48:28 INFO: step 2400: train_loss = 3.659402, dev_score = 0.6064
-2025-12-03 22:48:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:48:28 INFO: new model checkpoint saved.
-2025-12-03 22:48:30 INFO: Finished STEP 2420/50000, loss = 3.808647 (0.070 sec/batch), lr: 0.003000
-2025-12-03 22:48:31 INFO: Finished STEP 2440/50000, loss = 3.675776 (0.070 sec/batch), lr: 0.003000
-2025-12-03 22:48:33 INFO: Finished STEP 2460/50000, loss = 4.045568 (0.071 sec/batch), lr: 0.003000
-2025-12-03 22:48:34 INFO: Finished STEP 2480/50000, loss = 2.708247 (0.071 sec/batch), lr: 0.003000
-2025-12-03 22:48:36 INFO: Finished STEP 2500/50000, loss = 3.840647 (0.072 sec/batch), lr: 0.003000
-2025-12-03 22:48:36 INFO: Evaluating on dev set...
-2025-12-03 22:48:36 INFO: LAS	MLAS	BLEX
-2025-12-03 22:48:36 INFO: 63.61	54.97	57.93
-2025-12-03 22:48:36 INFO: step 2500: train_loss = 3.616523, dev_score = 0.6361
-2025-12-03 22:48:36 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 22:48:36 INFO: new best model saved.
-2025-12-03 22:48:37 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:48:37 INFO: new model checkpoint saved.
-2025-12-03 22:48:38 INFO: Finished STEP 2520/50000, loss = 3.329048 (0.075 sec/batch), lr: 0.003000
-2025-12-03 22:48:40 INFO: Finished STEP 2540/50000, loss = 3.042371 (0.066 sec/batch), lr: 0.003000
-2025-12-03 22:48:41 INFO: Finished STEP 2560/50000, loss = 3.319040 (0.070 sec/batch), lr: 0.003000
-2025-12-03 22:48:43 INFO: Finished STEP 2580/50000, loss = 4.341519 (0.056 sec/batch), lr: 0.003000
-2025-12-03 22:48:44 INFO: Finished STEP 2600/50000, loss = 5.309865 (0.070 sec/batch), lr: 0.003000
-2025-12-03 22:48:44 INFO: Evaluating on dev set...
-2025-12-03 22:48:44 INFO: LAS	MLAS	BLEX
-2025-12-03 22:48:44 INFO: 61.14	52.01	56.24
-2025-12-03 22:48:44 INFO: step 2600: train_loss = 3.623874, dev_score = 0.6114
-2025-12-03 22:48:45 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:48:45 INFO: new model checkpoint saved.
-2025-12-03 22:48:46 INFO: Finished STEP 2620/50000, loss = 2.560167 (0.066 sec/batch), lr: 0.003000
-2025-12-03 22:48:48 INFO: Finished STEP 2640/50000, loss = 3.262659 (0.066 sec/batch), lr: 0.003000
-2025-12-03 22:48:49 INFO: Finished STEP 2660/50000, loss = 2.611564 (0.072 sec/batch), lr: 0.003000
-2025-12-03 22:48:51 INFO: Finished STEP 2680/50000, loss = 3.953672 (0.070 sec/batch), lr: 0.003000
-2025-12-03 22:48:52 INFO: Finished STEP 2700/50000, loss = 3.726388 (0.072 sec/batch), lr: 0.003000
-2025-12-03 22:48:52 INFO: Evaluating on dev set...
-2025-12-03 22:48:53 INFO: LAS	MLAS	BLEX
-2025-12-03 22:48:53 INFO: 61.88	53.39	57.20
-2025-12-03 22:48:53 INFO: step 2700: train_loss = 3.740555, dev_score = 0.6188
-2025-12-03 22:48:53 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:48:53 INFO: new model checkpoint saved.
-2025-12-03 22:48:54 INFO: Finished STEP 2720/50000, loss = 3.536426 (0.066 sec/batch), lr: 0.003000
-2025-12-03 22:48:56 INFO: Finished STEP 2740/50000, loss = 4.492881 (0.072 sec/batch), lr: 0.003000
-2025-12-03 22:48:57 INFO: Finished STEP 2760/50000, loss = 3.437390 (0.073 sec/batch), lr: 0.003000
-2025-12-03 22:48:59 INFO: Finished STEP 2780/50000, loss = 3.811538 (0.065 sec/batch), lr: 0.003000
-2025-12-03 22:49:00 INFO: Finished STEP 2800/50000, loss = 2.615445 (0.069 sec/batch), lr: 0.003000
-2025-12-03 22:49:00 INFO: Evaluating on dev set...
-2025-12-03 22:49:01 INFO: LAS	MLAS	BLEX
-2025-12-03 22:49:01 INFO: 61.14	51.59	56.66
-2025-12-03 22:49:01 INFO: step 2800: train_loss = 3.751917, dev_score = 0.6114
-2025-12-03 22:49:01 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:49:01 INFO: new model checkpoint saved.
-2025-12-03 22:49:03 INFO: Finished STEP 2820/50000, loss = 3.525189 (0.076 sec/batch), lr: 0.003000
-2025-12-03 22:49:04 INFO: Finished STEP 2840/50000, loss = 3.113450 (0.066 sec/batch), lr: 0.003000
-2025-12-03 22:49:05 INFO: Finished STEP 2860/50000, loss = 2.488067 (0.066 sec/batch), lr: 0.003000
-2025-12-03 22:49:07 INFO: Finished STEP 2880/50000, loss = 2.702721 (0.071 sec/batch), lr: 0.003000
-2025-12-03 22:49:08 INFO: Finished STEP 2900/50000, loss = 3.248651 (0.068 sec/batch), lr: 0.003000
-2025-12-03 22:49:08 INFO: Evaluating on dev set...
-2025-12-03 22:49:09 INFO: LAS	MLAS	BLEX
-2025-12-03 22:49:09 INFO: 62.13	52.43	56.24
-2025-12-03 22:49:09 INFO: step 2900: train_loss = 3.761651, dev_score = 0.6213
-2025-12-03 22:49:09 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:49:09 INFO: new model checkpoint saved.
-2025-12-03 22:49:11 INFO: Finished STEP 2920/50000, loss = 3.129019 (0.072 sec/batch), lr: 0.003000
-2025-12-03 22:49:12 INFO: Finished STEP 2940/50000, loss = 3.579517 (0.067 sec/batch), lr: 0.003000
-2025-12-03 22:49:13 INFO: Finished STEP 2960/50000, loss = 3.505895 (0.067 sec/batch), lr: 0.003000
-2025-12-03 22:49:15 INFO: Finished STEP 2980/50000, loss = 2.193599 (0.073 sec/batch), lr: 0.003000
-2025-12-03 22:49:16 INFO: Finished STEP 3000/50000, loss = 1.882619 (0.064 sec/batch), lr: 0.003000
-2025-12-03 22:49:16 INFO: Evaluating on dev set...
-2025-12-03 22:49:17 INFO: LAS	MLAS	BLEX
-2025-12-03 22:49:17 INFO: 61.63	51.16	55.81
-2025-12-03 22:49:17 INFO: step 3000: train_loss = 4.014063, dev_score = 0.6163
-2025-12-03 22:49:17 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:49:17 INFO: new model checkpoint saved.
-2025-12-03 22:49:19 INFO: Finished STEP 3020/50000, loss = 5.977424 (0.058 sec/batch), lr: 0.003000
-2025-12-03 22:49:20 INFO: Finished STEP 3040/50000, loss = 3.319470 (0.063 sec/batch), lr: 0.003000
-2025-12-03 22:49:21 INFO: Finished STEP 3060/50000, loss = 3.211635 (0.069 sec/batch), lr: 0.003000
-2025-12-03 22:49:23 INFO: Finished STEP 3080/50000, loss = 3.183325 (0.065 sec/batch), lr: 0.003000
-2025-12-03 22:49:24 INFO: Finished STEP 3100/50000, loss = 2.788619 (0.069 sec/batch), lr: 0.003000
-2025-12-03 22:49:24 INFO: Evaluating on dev set...
-2025-12-03 22:49:25 INFO: LAS	MLAS	BLEX
-2025-12-03 22:49:25 INFO: 60.89	51.05	56.54
-2025-12-03 22:49:25 INFO: step 3100: train_loss = 3.763229, dev_score = 0.6089
-2025-12-03 22:49:25 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:49:25 INFO: new model checkpoint saved.
-2025-12-03 22:49:27 INFO: Finished STEP 3120/50000, loss = 4.927651 (0.063 sec/batch), lr: 0.003000
-2025-12-03 22:49:28 INFO: Finished STEP 3140/50000, loss = 3.578332 (0.068 sec/batch), lr: 0.003000
-2025-12-03 22:49:29 INFO: Finished STEP 3160/50000, loss = 3.173575 (0.076 sec/batch), lr: 0.003000
-2025-12-03 22:49:31 INFO: Finished STEP 3180/50000, loss = 4.860454 (0.075 sec/batch), lr: 0.003000
-2025-12-03 22:49:32 INFO: Finished STEP 3200/50000, loss = 1.778136 (0.065 sec/batch), lr: 0.003000
-2025-12-03 22:49:32 INFO: Evaluating on dev set...
-2025-12-03 22:49:33 INFO: LAS	MLAS	BLEX
-2025-12-03 22:49:33 INFO: 60.15	50.74	54.12
-2025-12-03 22:49:33 INFO: step 3200: train_loss = 3.791774, dev_score = 0.6015
-2025-12-03 22:49:33 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:49:33 INFO: new model checkpoint saved.
-2025-12-03 22:49:35 INFO: Finished STEP 3220/50000, loss = 8.492014 (0.070 sec/batch), lr: 0.003000
-2025-12-03 22:49:36 INFO: Finished STEP 3240/50000, loss = 3.045755 (0.064 sec/batch), lr: 0.003000
-2025-12-03 22:49:37 INFO: Finished STEP 3260/50000, loss = 4.122291 (0.076 sec/batch), lr: 0.003000
-2025-12-03 22:49:39 INFO: Finished STEP 3280/50000, loss = 3.458145 (0.063 sec/batch), lr: 0.003000
-2025-12-03 22:49:40 INFO: Finished STEP 3300/50000, loss = 2.681028 (0.064 sec/batch), lr: 0.003000
-2025-12-03 22:49:40 INFO: Evaluating on dev set...
-2025-12-03 22:49:41 INFO: LAS	MLAS	BLEX
-2025-12-03 22:49:41 INFO: 61.88	51.59	55.39
-2025-12-03 22:49:41 INFO: step 3300: train_loss = 3.750064, dev_score = 0.6188
-2025-12-03 22:49:41 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:49:41 INFO: new model checkpoint saved.
-2025-12-03 22:49:42 INFO: Finished STEP 3320/50000, loss = 3.499715 (0.066 sec/batch), lr: 0.003000
-2025-12-03 22:49:44 INFO: Finished STEP 3340/50000, loss = 3.765631 (0.064 sec/batch), lr: 0.003000
-2025-12-03 22:49:45 INFO: Finished STEP 3360/50000, loss = 2.908885 (0.064 sec/batch), lr: 0.003000
-2025-12-03 22:49:47 INFO: Finished STEP 3380/50000, loss = 4.147782 (0.067 sec/batch), lr: 0.003000
-2025-12-03 22:49:48 INFO: Finished STEP 3400/50000, loss = 4.854831 (0.073 sec/batch), lr: 0.003000
-2025-12-03 22:49:48 INFO: Evaluating on dev set...
-2025-12-03 22:49:48 INFO: LAS	MLAS	BLEX
-2025-12-03 22:49:48 INFO: 61.39	51.59	54.55
-2025-12-03 22:49:48 INFO: step 3400: train_loss = 3.827870, dev_score = 0.6139
-2025-12-03 22:49:49 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 22:49:49 INFO: new model checkpoint saved.
-2025-12-03 22:49:50 INFO: Finished STEP 3420/50000, loss = 3.672020 (0.066 sec/batch), lr: 0.003000
-2025-12-03 22:49:52 INFO: Finished STEP 3440/50000, loss = 3.315593 (0.064 sec/batch), lr: 0.003000
-2025-12-03 22:49:53 INFO: Finished STEP 3460/50000, loss = 2.334443 (0.064 sec/batch), lr: 0.003000
-2025-12-03 22:49:54 INFO: Finished STEP 3480/50000, loss = 4.692723 (0.061 sec/batch), lr: 0.003000
-2025-12-03 22:49:56 INFO: Finished STEP 3500/50000, loss = 2.920178 (0.064 sec/batch), lr: 0.003000
-2025-12-03 22:49:56 INFO: Evaluating on dev set...
-2025-12-03 22:49:56 INFO: LAS	MLAS	BLEX
-2025-12-03 22:49:56 INFO: 60.40	50.74	53.70
-2025-12-03 22:49:56 INFO: step 3500: train_loss = 3.750002, dev_score = 0.6040
-2025-12-03 22:49:56 INFO: Training ended with 3500 steps.
-2025-12-03 22:49:56 INFO: Best dev F1 = 63.61, at iteration = 2500
-2025-12-03 22:49:57 INFO: Running dev depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
-2025-12-03 22:49:57 INFO: Running parser in predict mode
-2025-12-03 22:49:57 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 22:49:59 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 22:49:59 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 22:49:59 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 22:49:59 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 22:50:00 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 22:50:00 INFO: Loading data with batch size 32...
-2025-12-03 22:50:00 DEBUG: 9 batches created.
-2025-12-03 22:50:00 INFO: F1 scores for each dependency:
-  Note that unlabeled attachment errors hurt the labeled attachment scores
-         acl: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
-   acl:relcl: p 0.2222 r 0.2857 f1 0.2500 (7 actual)
-       advcl: p 0.1667 r 0.2000 f1 0.1818 (5 actual)
-      advmod: p 0.4828 r 0.5600 f1 0.5185 (25 actual)
-        amod: p 0.8889 r 0.7742 f1 0.8276 (31 actual)
-       appos: p 1.0000 r 0.5000 f1 0.6667 (4 actual)
-         aux: p 0.8182 r 0.8182 f1 0.8182 (11 actual)
-        case: p 0.9444 r 0.9107 f1 0.9273 (56 actual)
-          cc: p 0.6923 r 0.6923 f1 0.6923 (13 actual)
-       ccomp: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-compound:prt: p 0.0000 r 0.0000 f1 0.0000 (0 actual)
-        conj: p 0.3750 r 0.5000 f1 0.4286 (12 actual)
-         cop: p 0.5000 r 0.3333 f1 0.4000 (3 actual)
-       csubj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-         det: p 0.8696 r 0.9091 f1 0.8889 (22 actual)
-        expl: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-        iobj: p 0.3333 r 0.5000 f1 0.4000 (2 actual)
-        mark: p 0.6364 r 0.5833 f1 0.6087 (12 actual)
-        nmod: p 0.4118 r 0.4667 f1 0.4375 (15 actual)
-   nmod:poss: p 1.0000 r 0.8947 f1 0.9444 (19 actual)
-       nsubj: p 0.5200 r 0.7647 f1 0.6190 (17 actual)
-  nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
-         obj: p 0.5769 r 0.6818 f1 0.6250 (22 actual)
-         obl: p 0.5714 r 0.5854 f1 0.5783 (41 actual)
-   obl:agent: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-      orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-   parataxis: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
-       punct: p 0.4423 r 0.4423 f1 0.4423 (52 actual)
-        root: p 0.5556 r 0.5556 f1 0.5556 (9 actual)
-       xcomp: p 0.6667 r 0.2500 f1 0.3636 (8 actual)
-2025-12-03 22:50:00 INFO: LAS	MLAS	BLEX
-2025-12-03 22:50:00 INFO: 63.61	54.97	57.93
-2025-12-03 22:50:00 INFO: Parser score:
-2025-12-03 22:50:00 INFO: sv_diachronic 63.61
-2025-12-03 22:50:00 INFO: Finished running dev set on
-UD_Swedish-diachronic
-  UAS   LAS  CLAS  MLAS  BLEX
-70.54 63.61 57.93 54.97 57.93
-2025-12-03 22:50:00 INFO: Running test depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
-2025-12-03 22:50:00 INFO: Running parser in predict mode
-2025-12-03 22:50:00 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 22:50:02 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 22:50:02 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 22:50:02 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 22:50:02 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 22:50:02 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 22:50:02 INFO: Loading data with batch size 32...
-2025-12-03 22:50:02 DEBUG: 93 batches created.
-2025-12-03 22:50:07 INFO: F1 scores for each dependency:
-  Note that unlabeled attachment errors hurt the labeled attachment scores
-         acl: p 0.3333 r 0.0312 f1 0.0571 (32 actual)
-   acl:cleft: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-   acl:relcl: p 0.3061 r 0.2000 f1 0.2419 (75 actual)
-       advcl: p 0.0893 r 0.1667 f1 0.1163 (60 actual)
- advcl:relcl: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-      advmod: p 0.5745 r 0.5896 f1 0.5820 (268 actual)
-        amod: p 0.8139 r 0.8174 f1 0.8156 (230 actual)
-       appos: p 0.0000 r 0.0000 f1 0.0000 (13 actual)
-         aux: p 0.8554 r 0.8452 f1 0.8503 (84 actual)
-    aux:pass: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-        case: p 0.8661 r 0.8150 f1 0.8398 (373 actual)
-          cc: p 0.6474 r 0.6516 f1 0.6495 (155 actual)
-       ccomp: p 0.0000 r 0.0000 f1 0.0000 (35 actual)
-compound:prt: p 0.6800 r 0.8095 f1 0.7391 (21 actual)
-        conj: p 0.2938 r 0.2975 f1 0.2956 (158 actual)
-         cop: p 0.7188 r 0.5000 f1 0.5897 (46 actual)
-       csubj: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-         dep: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-         det: p 0.8308 r 0.8029 f1 0.8166 (208 actual)
-   discourse: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
-  dislocated: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-        expl: p 0.5000 r 0.0909 f1 0.1538 (11 actual)
-     expl:pv: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-       fixed: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
-        flat: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-   flat:name: p 0.0000 r 0.0000 f1 0.0000 (12 actual)
-    goeswith: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-        iobj: p 0.1935 r 0.4286 f1 0.2667 (14 actual)
-        mark: p 0.6624 r 0.6797 f1 0.6710 (153 actual)
-        nmod: p 0.2661 r 0.2843 f1 0.2749 (102 actual)
-   nmod:poss: p 0.8865 r 0.8803 f1 0.8834 (142 actual)
-       nsubj: p 0.5385 r 0.6750 f1 0.5990 (280 actual)
-  nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (25 actual)
-      nummod: p 0.8000 r 0.8000 f1 0.8000 (10 actual)
-         obj: p 0.5556 r 0.6011 f1 0.5774 (183 actual)
-         obl: p 0.5030 r 0.5935 f1 0.5446 (278 actual)
-   obl:agent: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-      orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-   parataxis: p 0.1000 r 0.2222 f1 0.1379 (18 actual)
-       punct: p 0.4685 r 0.4729 f1 0.4707 (425 actual)
-  reparandum: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-        root: p 0.5657 r 0.5657 f1 0.5657 (99 actual)
-    vocative: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
-       xcomp: p 0.5000 r 0.3467 f1 0.4094 (75 actual)
-2025-12-03 22:50:07 INFO: LAS	MLAS	BLEX
-2025-12-03 22:50:07 INFO: 59.06	49.85	53.26
-2025-12-03 22:50:07 INFO: Parser score:
-2025-12-03 22:50:07 INFO: sv_diachronic 59.06
-2025-12-03 22:50:07 INFO: Finished running test set on
-UD_Swedish-diachronic
-  UAS   LAS  CLAS  MLAS  BLEX
-68.07 59.06 53.26 49.85 53.26
-DONE.
-Full log saved to: logs/log_conll17.pt_sv_diachron_20251203_223822.txt
-Symlink updated: logs/latest.txt → log_conll17.pt_sv_diachron_20251203_223822.txt

logs/log_conll17.pt_sv_diachron_20251212_145854.txt

@@ -1,161 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_diachron_20251212_145854.txt ===
-Language codes: sv diachron
-Using pretrained model: conll17.pt
-
-Running: python prepare-train-val-test.py sv diachron
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec991-spf148.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec988-spf145.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec1102-spf259.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-letter141673-Stalhammar.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec1033-spf190.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec25-Runius.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec987-spf144.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec1063-spf220.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Including DigPhil MACHINE in TRAIN (minus gold)…
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
-Cleaning TRAIN...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-[REMOVED] sent_id=6  ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
-[REMOVED] sent_id=7_8  ERRORS=['Multiple roots found: [5, 10]']
-[REMOVED] sent_id=30_31  ERRORS=['Multiple roots found: [3, 18]']
-[REMOVED] sent_id=35  ERRORS=['Line 36: Invalid token ID or head']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [7, 20]']
-[REMOVED] sent_id=8_9  ERRORS=['Multiple roots found: [24, 57]']
-[REMOVED] sent_id=12_13  ERRORS=['Multiple roots found: [11, 16]']
-[REMOVED] sent_id=124_split2  ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
-[REMOVED] sent_id=396  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=416  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=589  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=909  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=912  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=3_split1  ERRORS=['Multiple roots found: [4, 15, 17]']
-[REMOVED] sent_id=3_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
-[REMOVED] sent_id=3_4  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=5_6  ERRORS=['Multiple roots found: [3, 24]']
-[REMOVED] sent_id=11_12_13  ERRORS=['Multiple roots found: [5, 17, 25]']
-[REMOVED] sent_id=119  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=188  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=223  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=268  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=325  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=388  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=399  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=475  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=505  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=520  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=562  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=711  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=731  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=867  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=884  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=923  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=939  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1086  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1251  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1345  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1459  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1656  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=87_88  ERRORS=['Multiple roots found: [3, 6]']
-[REMOVED] sent_id=65_split2_66_split2  ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
-[REMOVED] sent_id=25  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=136  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=208  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=230  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=245  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=276  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=320  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=366  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=519  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=569  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=50_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
-[REMOVED] sent_id=53_54  ERRORS=['Multiple roots found: [27, 91]']
-[REMOVED] sent_id=55_56_57  ERRORS=['Multiple roots found: [2, 4, 13]']
-[REMOVED] sent_id=17_split1  ERRORS=['Multiple roots found: [2, 14, 17]']
-[REMOVED] sent_id=17_split2  ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
-[REMOVED] sent_id=19_split1  ERRORS=['Multiple roots found: [3, 31]']
-Cleaning DEV...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Cleaning TEST...
-Writing TRAIN → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (60334 valid sentences)
-Writing DEV → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (9 valid sentences)
-Writing TEST → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (99 valid sentences)
-Done.
-Sourcing scripts/config_alvis.sh
-Running stanza dataset preparation…
-2025-12-12 14:59:02 INFO: Datasets program called with:
-/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-12 14:59:02 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
-
-2025-12-12 14:59:02 INFO: Downloaded file to /cephyr/users/cleland/Alvis/stanza_resources/resources.json
-2025-12-12 14:59:02 DEBUG: Processing parameter "processors"...
-2025-12-12 14:59:02 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
-2025-12-12 14:59:02 INFO: Downloading these customized packages for language: sv (Swedish)...
-=======================
-| Processor | Package |
------------------------
-=======================
-
-2025-12-12 14:59:02 INFO: Finished downloading models and saved to /cephyr/users/cleland/Alvis/stanza_resources
-2025-12-12 14:59:02 INFO: Using tagger model in /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt for sv_diachronic
-2025-12-12 14:59:02 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2025-12-12 14:59:02 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-Augmented 189 quotes: Counter({'""': 26, '「」': 23, '„”': 23, '»«': 22, '″″': 18, '””': 18, '““': 17, '《》': 17, '„“': 13, '«»': 12})
-2025-12-12 14:59:05 INFO: Running tagger to retag /local/tmp.5491708/tmpc2soyxjt/sv_diachronic.train.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5491708/tmpc2soyxjt/sv_diachronic.train.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu']
-2025-12-12 14:59:05 INFO: Running tagger in predict mode
-2025-12-12 14:59:05 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2025-12-12 14:59:07 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-12 14:59:07 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-12 14:59:07 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-12 14:59:07 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-12 14:59:08 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-12 14:59:11 INFO: Loading data with batch size 250...
-./make_new_model.sh: line 58: 3492354 Terminated              python -m stanza.utils.datasets.prepare_depparse_treebank UD_Swedish-diachronic --wordvec_pretrain_file "/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/${PRETRAINED_MODEL}"

logs/log_conll17.pt_sv_diachron_20260116_065352.txt

@@ -1,830 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_diachron_20260116_065352.txt ===
-Language codes: sv diachron
-Using pretrained model: conll17.pt
-
-Running: python prepare-train-val-test.py sv diachron
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec991-spf148.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec988-spf145.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec1102-spf259.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-letter141673-Stalhammar.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec1033-spf190.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec25-Runius.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec987-spf144.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec1063-spf220.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Including DigPhil MACHINE in TRAIN (minus gold)…
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
-Cleaning TRAIN...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-[REMOVED] sent_id=6  ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
-[REMOVED] sent_id=7_8  ERRORS=['Multiple roots found: [5, 10]']
-[REMOVED] sent_id=30_31  ERRORS=['Multiple roots found: [3, 18]']
-[REMOVED] sent_id=35  ERRORS=['Line 36: Invalid token ID or head']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [7, 20]']
-[REMOVED] sent_id=8_9  ERRORS=['Multiple roots found: [24, 57]']
-[REMOVED] sent_id=12_13  ERRORS=['Multiple roots found: [11, 16]']
-[REMOVED] sent_id=124_split2  ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
-[REMOVED] sent_id=396  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=416  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=589  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=909  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=912  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=3_split1  ERRORS=['Multiple roots found: [4, 15, 17]']
-[REMOVED] sent_id=3_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
-[REMOVED] sent_id=3_4  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=5_6  ERRORS=['Multiple roots found: [3, 24]']
-[REMOVED] sent_id=11_12_13  ERRORS=['Multiple roots found: [5, 17, 25]']
-[REMOVED] sent_id=119  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=188  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=223  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=268  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=325  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=388  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=399  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=475  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=505  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=520  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=562  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=711  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=731  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=867  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=884  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=923  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=939  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1086  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1251  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1345  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1459  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1656  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=87_88  ERRORS=['Multiple roots found: [3, 6]']
-[REMOVED] sent_id=65_split2_66_split2  ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
-[REMOVED] sent_id=25  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=136  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=208  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=230  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=245  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=276  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=320  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=366  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=519  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=569  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=50_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
-[REMOVED] sent_id=53_54  ERRORS=['Multiple roots found: [27, 91]']
-[REMOVED] sent_id=55_56_57  ERRORS=['Multiple roots found: [2, 4, 13]']
-[REMOVED] sent_id=17_split1  ERRORS=['Multiple roots found: [2, 14, 17]']
-[REMOVED] sent_id=17_split2  ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
-[REMOVED] sent_id=19_split1  ERRORS=['Multiple roots found: [3, 31]']
-Cleaning DEV...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Cleaning TEST...
-Writing TRAIN → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (60334 valid sentences)
-Writing DEV → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (9 valid sentences)
-Writing TEST → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (99 valid sentences)
-Done.
-Sourcing scripts/config_alvis.sh
-Running stanza dataset preparation…
-2026-01-16 06:54:03 INFO: Datasets program called with:
-/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2026-01-16 06:54:03 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
-
-2026-01-16 06:54:03 INFO: Downloaded file to /cephyr/users/cleland/Alvis/stanza_resources/resources.json
-2026-01-16 06:54:03 DEBUG: Processing parameter "processors"...
-2026-01-16 06:54:03 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
-2026-01-16 06:54:03 INFO: Downloading these customized packages for language: sv (Swedish)...
-=======================
-| Processor | Package |
------------------------
-=======================
-
-2026-01-16 06:54:03 INFO: Finished downloading models and saved to /cephyr/users/cleland/Alvis/stanza_resources
-2026-01-16 06:54:03 INFO: Using tagger model in /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt for sv_diachronic
-2026-01-16 06:54:03 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2026-01-16 06:54:03 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-Augmented 189 quotes: Counter({'""': 26, '「」': 23, '„”': 23, '»«': 22, '″″': 18, '””': 18, '““': 17, '《》': 17, '„“': 13, '«»': 12})
-2026-01-16 06:54:06 INFO: Running tagger to retag /local/tmp.5658622/tmp9_y5_489/sv_diachronic.train.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5658622/tmp9_y5_489/sv_diachronic.train.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu']
-2026-01-16 06:54:06 INFO: Running tagger in predict mode
-2026-01-16 06:54:06 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2026-01-16 06:54:08 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2026-01-16 06:54:08 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 06:54:08 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2026-01-16 06:54:08 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 06:54:08 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2026-01-16 06:54:12 INFO: Loading data with batch size 250...
-2026-01-16 06:54:44 INFO: Start evaluation...
-2026-01-16 06:57:49 INFO: UPOS	XPOS	UFeats	AllTags
-2026-01-16 06:57:49 INFO: 99.37	87.38	98.56	87.09
-2026-01-16 06:57:49 INFO: POS Tagger score: sv_diachronic 87.09
-2026-01-16 06:57:50 INFO: Running tagger to retag /local/tmp.5658622/tmp9_y5_489/sv_diachronic.dev.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5658622/tmp9_y5_489/sv_diachronic.dev.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu']
-2026-01-16 06:57:50 INFO: Running tagger in predict mode
-2026-01-16 06:57:50 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2026-01-16 06:57:51 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2026-01-16 06:57:51 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 06:57:51 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2026-01-16 06:57:52 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 06:57:52 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2026-01-16 06:57:52 INFO: Loading data with batch size 250...
-2026-01-16 06:57:52 INFO: Start evaluation...
-2026-01-16 06:57:52 INFO: UPOS	XPOS	UFeats	AllTags
-2026-01-16 06:57:52 INFO: 93.32	90.84	93.32	85.64
-2026-01-16 06:57:52 INFO: POS Tagger score: sv_diachronic 85.64
-2026-01-16 06:57:52 INFO: Running tagger to retag /local/tmp.5658622/tmp9_y5_489/sv_diachronic.test.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5658622/tmp9_y5_489/sv_diachronic.test.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu']
-2026-01-16 06:57:52 INFO: Running tagger in predict mode
-2026-01-16 06:57:52 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2026-01-16 06:57:54 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2026-01-16 06:57:54 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 06:57:54 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2026-01-16 06:57:54 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 06:57:54 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2026-01-16 06:57:54 INFO: Loading data with batch size 250...
-2026-01-16 06:57:54 INFO: Start evaluation...
-2026-01-16 06:57:55 INFO: UPOS	XPOS	UFeats	AllTags
-2026-01-16 06:57:55 INFO: 93.14	96.78	95.32	90.28
-2026-01-16 06:57:55 INFO: POS Tagger score: sv_diachronic 90.28
-Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /local/tmp.5658622/tmp9_y5_489/sv_diachronic.train.gold.conllu
-Swapped 'w1, w2' for 'w1 ,w2' 106 times
-Added 520 new sentences with asdf, zzzz -> asdf,zzzz
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /local/tmp.5658622/tmp9_y5_489/sv_diachronic.dev.gold.conllu
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /local/tmp.5658622/tmp9_y5_489/sv_diachronic.test.gold.conllu
-Running stanza dependency parser training…
-2026-01-16 06:58:08 INFO: Training program called with:
-/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt --batch_size 8 --dropout 0.33
-2026-01-16 06:58:08 DEBUG: UD_Swedish-diachronic: sv_diachronic
-2026-01-16 06:58:08 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2026-01-16 06:58:08 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2026-01-16 06:58:08 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_charlm_parser.pt does not exist, training new model
-2026-01-16 06:58:08 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2026-01-16 06:58:08 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2026-01-16 06:58:08 INFO: Running train depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--train_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--batch_size', '5000', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'train', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '8', '--dropout', '0.33']
-2026-01-16 06:58:08 INFO: Running parser in train mode
-2026-01-16 06:58:08 INFO: Using pretrained contextualized char embedding
-2026-01-16 06:58:08 INFO: Loading data with batch size 8...
-2026-01-16 06:58:14 INFO: Train File /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu, Data Size: 60854
-2026-01-16 06:58:14 INFO: Original data size: 60854
-2026-01-16 06:58:14 INFO: Augmented data size: 60854
-2026-01-16 06:58:29 WARNING: sv_diachronic is not a known dataset.  Examining the data to choose which xpos vocab to use
-2026-01-16 06:58:29 INFO: Original length = 60854
-2026-01-16 06:58:29 INFO: Filtered length = 60854
-2026-01-16 06:58:43 WARNING: Chose XPOSDescription(xpos_type=<XPOSType.XPOS: 1>, sep='|') for the xpos factory for sv_diachronic
-2026-01-16 06:58:48 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2026-01-16 06:58:59 DEBUG: 58067 batches created.
-2026-01-16 06:58:59 DEBUG: 9 batches created.
-2026-01-16 06:58:59 INFO: Training parser...
-2026-01-16 06:58:59 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 06:58:59 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2026-01-16 06:59:00 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 06:59:00 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2026-01-16 06:59:04 INFO: Finished STEP 20/50000, loss = 0.000000 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:05 INFO: Finished STEP 40/50000, loss = 0.000000 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:05 INFO: Finished STEP 60/50000, loss = 0.000000 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:06 INFO: Finished STEP 80/50000, loss = 0.000000 (0.032 sec/batch), lr: 0.003000
-2026-01-16 06:59:07 INFO: Finished STEP 100/50000, loss = 0.000000 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:07 INFO: Evaluating on dev set...
-2026-01-16 06:59:07 INFO: LAS	MLAS	BLEX
-2026-01-16 06:59:07 INFO: 0.50	0.31	0.63
-2026-01-16 06:59:07 INFO: step 100: train_loss = 192.073832, dev_score = 0.0050
-2026-01-16 06:59:07 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 06:59:07 INFO: new best model saved.
-2026-01-16 06:59:08 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 06:59:08 INFO: new model checkpoint saved.
-2026-01-16 06:59:08 INFO: Finished STEP 120/50000, loss = 0.000000 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:09 INFO: Finished STEP 140/50000, loss = 0.000000 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:10 INFO: Finished STEP 160/50000, loss = 0.000000 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:10 INFO: Finished STEP 180/50000, loss = 0.000000 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:11 INFO: Finished STEP 200/50000, loss = 0.000000 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:11 INFO: Evaluating on dev set...
-2026-01-16 06:59:11 INFO: LAS	MLAS	BLEX
-2026-01-16 06:59:11 INFO: 0.50	0.31	0.63
-2026-01-16 06:59:11 INFO: step 200: train_loss = 0.000000, dev_score = 0.0050
-2026-01-16 06:59:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 06:59:12 INFO: new best model saved.
-2026-01-16 06:59:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 06:59:12 INFO: new model checkpoint saved.
-2026-01-16 06:59:13 INFO: Finished STEP 220/50000, loss = 0.000000 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:13 INFO: Finished STEP 240/50000, loss = 0.000000 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:14 INFO: Finished STEP 260/50000, loss = 0.000000 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:15 INFO: Finished STEP 280/50000, loss = 0.000000 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:15 INFO: Finished STEP 300/50000, loss = 2.877054 (0.032 sec/batch), lr: 0.003000
-2026-01-16 06:59:15 INFO: Evaluating on dev set...
-2026-01-16 06:59:16 INFO: LAS	MLAS	BLEX
-2026-01-16 06:59:16 INFO: 1.49	0.00	0.00
-2026-01-16 06:59:16 INFO: step 300: train_loss = 32.398629, dev_score = 0.0149
-2026-01-16 06:59:16 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 06:59:16 INFO: new best model saved.
-2026-01-16 06:59:17 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 06:59:17 INFO: new model checkpoint saved.
-2026-01-16 06:59:17 INFO: Finished STEP 320/50000, loss = 4.388504 (0.032 sec/batch), lr: 0.003000
-2026-01-16 06:59:18 INFO: Finished STEP 340/50000, loss = 1.865119 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:18 INFO: Finished STEP 360/50000, loss = 2.029490 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:19 INFO: Finished STEP 380/50000, loss = 3.120337 (0.033 sec/batch), lr: 0.003000
-2026-01-16 06:59:20 INFO: Finished STEP 400/50000, loss = 0.527415 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:20 INFO: Evaluating on dev set...
-2026-01-16 06:59:20 INFO: LAS	MLAS	BLEX
-2026-01-16 06:59:20 INFO: 2.97	2.74	2.74
-2026-01-16 06:59:20 INFO: step 400: train_loss = 2.460383, dev_score = 0.0297
-2026-01-16 06:59:21 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 06:59:21 INFO: new best model saved.
-2026-01-16 06:59:21 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 06:59:21 INFO: new model checkpoint saved.
-2026-01-16 06:59:22 INFO: Finished STEP 420/50000, loss = 0.000224 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:22 INFO: Finished STEP 440/50000, loss = 1.204638 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:23 INFO: Finished STEP 460/50000, loss = 1.052635 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:24 INFO: Finished STEP 480/50000, loss = 2.311400 (0.032 sec/batch), lr: 0.003000
-2026-01-16 06:59:24 INFO: Finished STEP 500/50000, loss = 1.092946 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:24 INFO: Evaluating on dev set...
-2026-01-16 06:59:25 INFO: LAS	MLAS	BLEX
-2026-01-16 06:59:25 INFO: 2.48	0.70	1.41
-2026-01-16 06:59:25 INFO: step 500: train_loss = 1.338739, dev_score = 0.0248
-2026-01-16 06:59:25 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 06:59:25 INFO: new model checkpoint saved.
-2026-01-16 06:59:26 INFO: Finished STEP 520/50000, loss = 0.652456 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:27 INFO: Finished STEP 540/50000, loss = 0.147004 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:27 INFO: Finished STEP 560/50000, loss = 0.782623 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:28 INFO: Finished STEP 580/50000, loss = 0.076848 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:28 INFO: Finished STEP 600/50000, loss = 0.429073 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:28 INFO: Evaluating on dev set...
-2026-01-16 06:59:29 INFO: LAS	MLAS	BLEX
-2026-01-16 06:59:29 INFO: 3.71	2.21	2.65
-2026-01-16 06:59:29 INFO: step 600: train_loss = 1.109608, dev_score = 0.0371
-2026-01-16 06:59:29 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 06:59:29 INFO: new best model saved.
-2026-01-16 06:59:30 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 06:59:30 INFO: new model checkpoint saved.
-2026-01-16 06:59:30 INFO: Finished STEP 620/50000, loss = 0.636850 (0.033 sec/batch), lr: 0.003000
-2026-01-16 06:59:31 INFO: Finished STEP 640/50000, loss = 0.097528 (0.032 sec/batch), lr: 0.003000
-2026-01-16 06:59:32 INFO: Finished STEP 660/50000, loss = 0.328014 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:32 INFO: Finished STEP 680/50000, loss = 1.418195 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:33 INFO: Finished STEP 700/50000, loss = 0.000002 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:33 INFO: Evaluating on dev set...
-2026-01-16 06:59:34 INFO: LAS	MLAS	BLEX
-2026-01-16 06:59:34 INFO: 3.47	2.50	2.50
-2026-01-16 06:59:34 INFO: step 700: train_loss = 1.152763, dev_score = 0.0347
-2026-01-16 06:59:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 06:59:34 INFO: new model checkpoint saved.
-2026-01-16 06:59:35 INFO: Finished STEP 720/50000, loss = 3.007647 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:35 INFO: Finished STEP 740/50000, loss = 1.066953 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:36 INFO: Finished STEP 760/50000, loss = 1.151768 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:37 INFO: Finished STEP 780/50000, loss = 1.043154 (0.032 sec/batch), lr: 0.003000
-2026-01-16 06:59:37 INFO: Finished STEP 800/50000, loss = 0.000000 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:37 INFO: Evaluating on dev set...
-2026-01-16 06:59:38 INFO: LAS	MLAS	BLEX
-2026-01-16 06:59:38 INFO: 3.71	2.21	2.58
-2026-01-16 06:59:38 INFO: step 800: train_loss = 1.042035, dev_score = 0.0371
-2026-01-16 06:59:38 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 06:59:38 INFO: new best model saved.
-2026-01-16 06:59:38 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 06:59:38 INFO: new model checkpoint saved.
-2026-01-16 06:59:39 INFO: Finished STEP 820/50000, loss = 0.309987 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:40 INFO: Finished STEP 840/50000, loss = 0.046532 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:40 INFO: Finished STEP 860/50000, loss = 2.719592 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:41 INFO: Finished STEP 880/50000, loss = 1.216663 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:42 INFO: Finished STEP 900/50000, loss = 1.323724 (0.032 sec/batch), lr: 0.003000
-2026-01-16 06:59:42 INFO: Evaluating on dev set...
-2026-01-16 06:59:42 INFO: LAS	MLAS	BLEX
-2026-01-16 06:59:42 INFO: 3.71	2.71	3.05
-2026-01-16 06:59:42 INFO: step 900: train_loss = 1.123192, dev_score = 0.0371
-2026-01-16 06:59:42 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 06:59:42 INFO: new best model saved.
-2026-01-16 06:59:43 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 06:59:43 INFO: new model checkpoint saved.
-2026-01-16 06:59:43 INFO: Finished STEP 920/50000, loss = 0.152121 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:44 INFO: Finished STEP 940/50000, loss = 0.897826 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:45 INFO: Finished STEP 960/50000, loss = 0.000258 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:45 INFO: Finished STEP 980/50000, loss = 2.134546 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:46 INFO: Finished STEP 1000/50000, loss = 0.035686 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:46 INFO: Evaluating on dev set...
-2026-01-16 06:59:46 INFO: LAS	MLAS	BLEX
-2026-01-16 06:59:46 INFO: 3.71	2.49	2.85
-2026-01-16 06:59:46 INFO: step 1000: train_loss = 0.984753, dev_score = 0.0371
-2026-01-16 06:59:47 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 06:59:47 INFO: new best model saved.
-2026-01-16 06:59:47 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 06:59:47 INFO: new model checkpoint saved.
-2026-01-16 06:59:48 INFO: Finished STEP 1020/50000, loss = 1.837077 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:48 INFO: Finished STEP 1040/50000, loss = 0.000000 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:49 INFO: Finished STEP 1060/50000, loss = 2.368280 (0.030 sec/batch), lr: 0.003000
-2026-01-16 06:59:50 INFO: Finished STEP 1080/50000, loss = 0.469828 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:50 INFO: Finished STEP 1100/50000, loss = 0.000491 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:50 INFO: Evaluating on dev set...
-2026-01-16 06:59:51 INFO: LAS	MLAS	BLEX
-2026-01-16 06:59:51 INFO: 4.95	4.11	5.48
-2026-01-16 06:59:51 INFO: step 1100: train_loss = 1.174486, dev_score = 0.0495
-2026-01-16 06:59:51 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 06:59:51 INFO: new best model saved.
-2026-01-16 06:59:52 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 06:59:52 INFO: new model checkpoint saved.
-2026-01-16 06:59:52 INFO: Finished STEP 1120/50000, loss = 5.331606 (0.033 sec/batch), lr: 0.003000
-2026-01-16 06:59:53 INFO: Finished STEP 1140/50000, loss = 4.487067 (0.032 sec/batch), lr: 0.003000
-2026-01-16 06:59:54 INFO: Finished STEP 1160/50000, loss = 1.940725 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:54 INFO: Finished STEP 1180/50000, loss = 2.472967 (0.032 sec/batch), lr: 0.003000
-2026-01-16 06:59:55 INFO: Finished STEP 1200/50000, loss = 0.686162 (0.032 sec/batch), lr: 0.003000
-2026-01-16 06:59:55 INFO: Evaluating on dev set...
-2026-01-16 06:59:56 INFO: LAS	MLAS	BLEX
-2026-01-16 06:59:56 INFO: 6.68	4.08	5.10
-2026-01-16 06:59:56 INFO: step 1200: train_loss = 2.657063, dev_score = 0.0668
-2026-01-16 06:59:56 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 06:59:56 INFO: new best model saved.
-2026-01-16 06:59:56 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 06:59:56 INFO: new model checkpoint saved.
-2026-01-16 06:59:57 INFO: Finished STEP 1220/50000, loss = 0.828940 (0.031 sec/batch), lr: 0.003000
-2026-01-16 06:59:58 INFO: Finished STEP 1240/50000, loss = 1.596623 (0.032 sec/batch), lr: 0.003000
-2026-01-16 06:59:58 INFO: Finished STEP 1260/50000, loss = 6.153142 (0.032 sec/batch), lr: 0.003000
-2026-01-16 06:59:59 INFO: Finished STEP 1280/50000, loss = 1.988011 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:00 INFO: Finished STEP 1300/50000, loss = 3.000342 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:00 INFO: Evaluating on dev set...
-2026-01-16 07:00:00 INFO: LAS	MLAS	BLEX
-2026-01-16 07:00:00 INFO: 7.67	5.80	6.48
-2026-01-16 07:00:00 INFO: step 1300: train_loss = 1.683935, dev_score = 0.0767
-2026-01-16 07:00:00 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:00:00 INFO: new best model saved.
-2026-01-16 07:00:01 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:00:01 INFO: new model checkpoint saved.
-2026-01-16 07:00:02 INFO: Finished STEP 1320/50000, loss = 2.859746 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:02 INFO: Finished STEP 1340/50000, loss = 1.214685 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:03 INFO: Finished STEP 1360/50000, loss = 0.369348 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:04 INFO: Finished STEP 1380/50000, loss = 1.401275 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:04 INFO: Finished STEP 1400/50000, loss = 1.433550 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:04 INFO: Evaluating on dev set...
-2026-01-16 07:00:05 INFO: LAS	MLAS	BLEX
-2026-01-16 07:00:05 INFO: 9.65	4.51	5.64
-2026-01-16 07:00:05 INFO: step 1400: train_loss = 1.665584, dev_score = 0.0965
-2026-01-16 07:00:05 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:00:05 INFO: new best model saved.
-2026-01-16 07:00:05 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:00:05 INFO: new model checkpoint saved.
-2026-01-16 07:00:06 INFO: Finished STEP 1420/50000, loss = 1.668130 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:07 INFO: Finished STEP 1440/50000, loss = 0.647845 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:07 INFO: Finished STEP 1460/50000, loss = 1.816998 (0.034 sec/batch), lr: 0.003000
-2026-01-16 07:00:08 INFO: Finished STEP 1480/50000, loss = 1.645519 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:09 INFO: Finished STEP 1500/50000, loss = 3.368581 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:09 INFO: Evaluating on dev set...
-2026-01-16 07:00:10 INFO: LAS	MLAS	BLEX
-2026-01-16 07:00:10 INFO: 11.14	6.09	6.81
-2026-01-16 07:00:10 INFO: step 1500: train_loss = 1.689450, dev_score = 0.1114
-2026-01-16 07:00:10 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:00:10 INFO: new best model saved.
-2026-01-16 07:00:10 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:00:10 INFO: new model checkpoint saved.
-2026-01-16 07:00:11 INFO: Finished STEP 1520/50000, loss = 0.696790 (0.033 sec/batch), lr: 0.003000
-2026-01-16 07:00:12 INFO: Finished STEP 1540/50000, loss = 0.979858 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:12 INFO: Finished STEP 1560/50000, loss = 1.371460 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:13 INFO: Finished STEP 1580/50000, loss = 1.076014 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:14 INFO: Finished STEP 1600/50000, loss = 0.347749 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:14 INFO: Evaluating on dev set...
-2026-01-16 07:00:14 INFO: LAS	MLAS	BLEX
-2026-01-16 07:00:14 INFO: 18.07	4.21	5.35
-2026-01-16 07:00:14 INFO: step 1600: train_loss = 1.549157, dev_score = 0.1807
-2026-01-16 07:00:14 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:00:14 INFO: new best model saved.
-2026-01-16 07:00:15 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:00:15 INFO: new model checkpoint saved.
-2026-01-16 07:00:16 INFO: Finished STEP 1620/50000, loss = 0.799195 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:16 INFO: Finished STEP 1640/50000, loss = 1.727682 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:17 INFO: Finished STEP 1660/50000, loss = 2.875053 (0.033 sec/batch), lr: 0.003000
-2026-01-16 07:00:18 INFO: Finished STEP 1680/50000, loss = 2.254138 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:18 INFO: Finished STEP 1700/50000, loss = 0.754430 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:18 INFO: Evaluating on dev set...
-2026-01-16 07:00:19 INFO: LAS	MLAS	BLEX
-2026-01-16 07:00:19 INFO: 11.39	4.62	5.01
-2026-01-16 07:00:19 INFO: step 1700: train_loss = 1.452840, dev_score = 0.1139
-2026-01-16 07:00:19 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:00:19 INFO: new model checkpoint saved.
-2026-01-16 07:00:20 INFO: Finished STEP 1720/50000, loss = 0.672960 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:21 INFO: Finished STEP 1740/50000, loss = 1.758359 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:21 INFO: Finished STEP 1760/50000, loss = 0.112267 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:22 INFO: Finished STEP 1780/50000, loss = 2.130425 (0.035 sec/batch), lr: 0.003000
-2026-01-16 07:00:23 INFO: Finished STEP 1800/50000, loss = 0.867664 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:23 INFO: Evaluating on dev set...
-2026-01-16 07:00:23 INFO: LAS	MLAS	BLEX
-2026-01-16 07:00:23 INFO: 8.17	5.17	5.52
-2026-01-16 07:00:23 INFO: step 1800: train_loss = 1.566285, dev_score = 0.0817
-2026-01-16 07:00:24 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:00:24 INFO: new model checkpoint saved.
-2026-01-16 07:00:24 INFO: Finished STEP 1820/50000, loss = 0.918402 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:25 INFO: Finished STEP 1840/50000, loss = 0.602067 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:26 INFO: Finished STEP 1860/50000, loss = 4.889768 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:26 INFO: Finished STEP 1880/50000, loss = 3.249488 (0.037 sec/batch), lr: 0.003000
-2026-01-16 07:00:27 INFO: Finished STEP 1900/50000, loss = 1.829388 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:27 INFO: Evaluating on dev set...
-2026-01-16 07:00:28 INFO: LAS	MLAS	BLEX
-2026-01-16 07:00:28 INFO: 9.16	3.77	5.28
-2026-01-16 07:00:28 INFO: step 1900: train_loss = 1.531276, dev_score = 0.0916
-2026-01-16 07:00:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:00:28 INFO: new model checkpoint saved.
-2026-01-16 07:00:29 INFO: Finished STEP 1920/50000, loss = 1.377406 (0.033 sec/batch), lr: 0.003000
-2026-01-16 07:00:29 INFO: Finished STEP 1940/50000, loss = 0.690662 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:30 INFO: Finished STEP 1960/50000, loss = 0.488795 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:31 INFO: Finished STEP 1980/50000, loss = 1.489328 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:31 INFO: Finished STEP 2000/50000, loss = 0.265036 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:31 INFO: Evaluating on dev set...
-2026-01-16 07:00:32 INFO: LAS	MLAS	BLEX
-2026-01-16 07:00:32 INFO: 3.71	2.34	3.90
-2026-01-16 07:00:32 INFO: step 2000: train_loss = 1.577713, dev_score = 0.0371
-2026-01-16 07:00:32 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:00:32 INFO: new model checkpoint saved.
-2026-01-16 07:00:33 INFO: Finished STEP 2020/50000, loss = 1.118794 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:34 INFO: Finished STEP 2040/50000, loss = 1.678119 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:34 INFO: Finished STEP 2060/50000, loss = 0.566802 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:35 INFO: Finished STEP 2080/50000, loss = 1.566789 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:36 INFO: Finished STEP 2100/50000, loss = 1.108176 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:36 INFO: Evaluating on dev set...
-2026-01-16 07:00:36 INFO: LAS	MLAS	BLEX
-2026-01-16 07:00:36 INFO: 8.17	5.37	6.91
-2026-01-16 07:00:36 INFO: step 2100: train_loss = 1.353945, dev_score = 0.0817
-2026-01-16 07:00:37 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:00:37 INFO: new model checkpoint saved.
-2026-01-16 07:00:37 INFO: Finished STEP 2120/50000, loss = 0.896808 (0.033 sec/batch), lr: 0.003000
-2026-01-16 07:00:38 INFO: Finished STEP 2140/50000, loss = 3.207397 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:39 INFO: Finished STEP 2160/50000, loss = 0.004755 (0.037 sec/batch), lr: 0.003000
-2026-01-16 07:00:39 INFO: Finished STEP 2180/50000, loss = 0.760959 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:40 INFO: Finished STEP 2200/50000, loss = 2.696491 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:40 INFO: Evaluating on dev set...
-2026-01-16 07:00:41 INFO: LAS	MLAS	BLEX
-2026-01-16 07:00:41 INFO: 7.67	1.83	2.19
-2026-01-16 07:00:41 INFO: step 2200: train_loss = 1.574935, dev_score = 0.0767
-2026-01-16 07:00:41 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:00:41 INFO: new model checkpoint saved.
-2026-01-16 07:00:42 INFO: Finished STEP 2220/50000, loss = 2.683285 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:42 INFO: Finished STEP 2240/50000, loss = 2.951740 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:43 INFO: Finished STEP 2260/50000, loss = 4.218589 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:44 INFO: Finished STEP 2280/50000, loss = 1.695766 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:44 INFO: Finished STEP 2300/50000, loss = 2.362932 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:44 INFO: Evaluating on dev set...
-2026-01-16 07:00:45 INFO: LAS	MLAS	BLEX
-2026-01-16 07:00:45 INFO: 12.62	3.67	4.48
-2026-01-16 07:00:45 INFO: step 2300: train_loss = 3.007224, dev_score = 0.1262
-2026-01-16 07:00:46 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:00:46 INFO: new model checkpoint saved.
-2026-01-16 07:00:46 INFO: Finished STEP 2320/50000, loss = 2.829835 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:47 INFO: Finished STEP 2340/50000, loss = 1.246457 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:47 INFO: Finished STEP 2360/50000, loss = 1.362900 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:48 INFO: Finished STEP 2380/50000, loss = 1.726964 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:49 INFO: Finished STEP 2400/50000, loss = 1.644795 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:49 INFO: Evaluating on dev set...
-2026-01-16 07:00:49 INFO: LAS	MLAS	BLEX
-2026-01-16 07:00:49 INFO: 11.14	6.71	7.77
-2026-01-16 07:00:49 INFO: step 2400: train_loss = 2.391629, dev_score = 0.1114
-2026-01-16 07:00:50 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:00:50 INFO: new model checkpoint saved.
-2026-01-16 07:00:50 INFO: Finished STEP 2420/50000, loss = 1.448694 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:00:51 INFO: Finished STEP 2440/50000, loss = 0.224588 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:00:52 INFO: Finished STEP 2460/50000, loss = 1.348104 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:52 INFO: Finished STEP 2480/50000, loss = 1.120497 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:53 INFO: Finished STEP 2500/50000, loss = 2.224450 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:53 INFO: Evaluating on dev set...
-2026-01-16 07:00:53 INFO: LAS	MLAS	BLEX
-2026-01-16 07:00:53 INFO: 18.56	6.11	7.63
-2026-01-16 07:00:53 INFO: step 2500: train_loss = 2.279046, dev_score = 0.1856
-2026-01-16 07:00:54 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:00:54 INFO: new best model saved.
-2026-01-16 07:00:54 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:00:54 INFO: new model checkpoint saved.
-2026-01-16 07:00:55 INFO: Finished STEP 2520/50000, loss = 4.519064 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:00:55 INFO: Finished STEP 2540/50000, loss = 3.034460 (0.034 sec/batch), lr: 0.003000
-2026-01-16 07:00:56 INFO: Finished STEP 2560/50000, loss = 2.673581 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:57 INFO: Finished STEP 2580/50000, loss = 0.801564 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:00:57 INFO: Finished STEP 2600/50000, loss = 0.521267 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:00:57 INFO: Evaluating on dev set...
-2026-01-16 07:00:58 INFO: LAS	MLAS	BLEX
-2026-01-16 07:00:58 INFO: 18.56	6.01	7.21
-2026-01-16 07:00:58 INFO: step 2600: train_loss = 2.528788, dev_score = 0.1856
-2026-01-16 07:00:58 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:00:58 INFO: new best model saved.
-2026-01-16 07:00:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:00:59 INFO: new model checkpoint saved.
-2026-01-16 07:00:59 INFO: Finished STEP 2620/50000, loss = 1.897941 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:00 INFO: Finished STEP 2640/50000, loss = 2.104889 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:00 INFO: Finished STEP 2660/50000, loss = 2.285084 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:01 INFO: Finished STEP 2680/50000, loss = 2.218475 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:02 INFO: Finished STEP 2700/50000, loss = 6.859745 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:02 INFO: Evaluating on dev set...
-2026-01-16 07:01:02 INFO: LAS	MLAS	BLEX
-2026-01-16 07:01:02 INFO: 19.55	6.19	8.12
-2026-01-16 07:01:02 INFO: step 2700: train_loss = 2.664627, dev_score = 0.1955
-2026-01-16 07:01:03 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:01:03 INFO: new best model saved.
-2026-01-16 07:01:03 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:01:03 INFO: new model checkpoint saved.
-2026-01-16 07:01:04 INFO: Finished STEP 2720/50000, loss = 2.067003 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:04 INFO: Finished STEP 2740/50000, loss = 1.080532 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:05 INFO: Finished STEP 2760/50000, loss = 4.616343 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:06 INFO: Finished STEP 2780/50000, loss = 1.479079 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:06 INFO: Finished STEP 2800/50000, loss = 1.063658 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:06 INFO: Evaluating on dev set...
-2026-01-16 07:01:07 INFO: LAS	MLAS	BLEX
-2026-01-16 07:01:07 INFO: 17.82	4.66	6.21
-2026-01-16 07:01:07 INFO: step 2800: train_loss = 2.274723, dev_score = 0.1782
-2026-01-16 07:01:08 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:01:08 INFO: new model checkpoint saved.
-2026-01-16 07:01:08 INFO: Finished STEP 2820/50000, loss = 2.248970 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:09 INFO: Finished STEP 2840/50000, loss = 3.309697 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:10 INFO: Finished STEP 2860/50000, loss = 0.586270 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:10 INFO: Finished STEP 2880/50000, loss = 1.731944 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:11 INFO: Finished STEP 2900/50000, loss = 2.552610 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:11 INFO: Evaluating on dev set...
-2026-01-16 07:01:11 INFO: LAS	MLAS	BLEX
-2026-01-16 07:01:11 INFO: 17.33	5.02	6.95
-2026-01-16 07:01:11 INFO: step 2900: train_loss = 2.368524, dev_score = 0.1733
-2026-01-16 07:01:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:01:12 INFO: new model checkpoint saved.
-2026-01-16 07:01:13 INFO: Finished STEP 2920/50000, loss = 3.126576 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:13 INFO: Finished STEP 2940/50000, loss = 1.741402 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:14 INFO: Finished STEP 2960/50000, loss = 3.777675 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:15 INFO: Finished STEP 2980/50000, loss = 1.370117 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:15 INFO: Finished STEP 3000/50000, loss = 3.574947 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:15 INFO: Evaluating on dev set...
-2026-01-16 07:01:16 INFO: LAS	MLAS	BLEX
-2026-01-16 07:01:16 INFO: 17.82	6.94	8.48
-2026-01-16 07:01:16 INFO: step 3000: train_loss = 2.539027, dev_score = 0.1782
-2026-01-16 07:01:16 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:01:16 INFO: new model checkpoint saved.
-2026-01-16 07:01:17 INFO: Finished STEP 3020/50000, loss = 1.453705 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:18 INFO: Finished STEP 3040/50000, loss = 6.184045 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:18 INFO: Finished STEP 3060/50000, loss = 4.618065 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:19 INFO: Finished STEP 3080/50000, loss = 2.472624 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:20 INFO: Finished STEP 3100/50000, loss = 2.051110 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:20 INFO: Evaluating on dev set...
-2026-01-16 07:01:20 INFO: LAS	MLAS	BLEX
-2026-01-16 07:01:20 INFO: 23.51	9.39	13.88
-2026-01-16 07:01:20 INFO: step 3100: train_loss = 2.557801, dev_score = 0.2351
-2026-01-16 07:01:20 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:01:20 INFO: new best model saved.
-2026-01-16 07:01:21 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:01:21 INFO: new model checkpoint saved.
-2026-01-16 07:01:21 INFO: Finished STEP 3120/50000, loss = 1.929233 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:22 INFO: Finished STEP 3140/50000, loss = 1.142895 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:23 INFO: Finished STEP 3160/50000, loss = 2.250463 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:23 INFO: Finished STEP 3180/50000, loss = 1.848980 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:24 INFO: Finished STEP 3200/50000, loss = 1.592728 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:24 INFO: Evaluating on dev set...
-2026-01-16 07:01:24 INFO: LAS	MLAS	BLEX
-2026-01-16 07:01:24 INFO: 14.11	7.38	10.10
-2026-01-16 07:01:24 INFO: step 3200: train_loss = 2.549193, dev_score = 0.1411
-2026-01-16 07:01:25 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:01:25 INFO: new model checkpoint saved.
-2026-01-16 07:01:26 INFO: Finished STEP 3220/50000, loss = 1.161050 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:26 INFO: Finished STEP 3240/50000, loss = 2.996624 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:27 INFO: Finished STEP 3260/50000, loss = 1.401396 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:28 INFO: Finished STEP 3280/50000, loss = 3.537535 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:28 INFO: Finished STEP 3300/50000, loss = 2.966022 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:28 INFO: Evaluating on dev set...
-2026-01-16 07:01:29 INFO: LAS	MLAS	BLEX
-2026-01-16 07:01:29 INFO: 21.78	6.75	7.54
-2026-01-16 07:01:29 INFO: step 3300: train_loss = 2.125264, dev_score = 0.2178
-2026-01-16 07:01:29 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:01:29 INFO: new model checkpoint saved.
-2026-01-16 07:01:30 INFO: Finished STEP 3320/50000, loss = 1.093041 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:30 INFO: Finished STEP 3340/50000, loss = 4.219791 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:31 INFO: Finished STEP 3360/50000, loss = 1.543054 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:32 INFO: Finished STEP 3380/50000, loss = 1.405899 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:32 INFO: Finished STEP 3400/50000, loss = 1.838370 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:32 INFO: Evaluating on dev set...
-2026-01-16 07:01:33 INFO: LAS	MLAS	BLEX
-2026-01-16 07:01:33 INFO: 11.39	5.43	9.69
-2026-01-16 07:01:33 INFO: step 3400: train_loss = 2.268760, dev_score = 0.1139
-2026-01-16 07:01:33 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:01:33 INFO: new model checkpoint saved.
-2026-01-16 07:01:34 INFO: Finished STEP 3420/50000, loss = 2.768859 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:35 INFO: Finished STEP 3440/50000, loss = 1.197934 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:35 INFO: Finished STEP 3460/50000, loss = 1.991076 (0.034 sec/batch), lr: 0.003000
-2026-01-16 07:01:36 INFO: Finished STEP 3480/50000, loss = 1.895437 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:37 INFO: Finished STEP 3500/50000, loss = 1.187534 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:37 INFO: Evaluating on dev set...
-2026-01-16 07:01:37 INFO: LAS	MLAS	BLEX
-2026-01-16 07:01:37 INFO: 19.06	10.51	15.76
-2026-01-16 07:01:37 INFO: step 3500: train_loss = 2.106533, dev_score = 0.1906
-2026-01-16 07:01:38 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:01:38 INFO: new model checkpoint saved.
-2026-01-16 07:01:38 INFO: Finished STEP 3520/50000, loss = 1.585278 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:01:39 INFO: Finished STEP 3540/50000, loss = 1.423405 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:39 INFO: Finished STEP 3560/50000, loss = 0.421177 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:40 INFO: Finished STEP 3580/50000, loss = 1.731428 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:41 INFO: Finished STEP 3600/50000, loss = 0.526926 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:41 INFO: Evaluating on dev set...
-2026-01-16 07:01:41 INFO: LAS	MLAS	BLEX
-2026-01-16 07:01:41 INFO: 19.55	8.20	10.55
-2026-01-16 07:01:41 INFO: step 3600: train_loss = 2.619668, dev_score = 0.1955
-2026-01-16 07:01:42 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:01:42 INFO: new model checkpoint saved.
-2026-01-16 07:01:42 INFO: Finished STEP 3620/50000, loss = 0.616525 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:43 INFO: Finished STEP 3640/50000, loss = 1.884155 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:44 INFO: Finished STEP 3660/50000, loss = 2.431090 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:44 INFO: Finished STEP 3680/50000, loss = 4.406909 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:01:45 INFO: Finished STEP 3700/50000, loss = 1.068943 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:45 INFO: Evaluating on dev set...
-2026-01-16 07:01:45 INFO: LAS	MLAS	BLEX
-2026-01-16 07:01:45 INFO: 10.15	6.90	9.20
-2026-01-16 07:01:45 INFO: step 3700: train_loss = 2.132062, dev_score = 0.1015
-2026-01-16 07:01:46 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:01:46 INFO: new model checkpoint saved.
-2026-01-16 07:01:46 INFO: Finished STEP 3720/50000, loss = 1.305085 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:47 INFO: Finished STEP 3740/50000, loss = 0.134945 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:48 INFO: Finished STEP 3760/50000, loss = 0.634447 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:48 INFO: Finished STEP 3780/50000, loss = 1.520501 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:49 INFO: Finished STEP 3800/50000, loss = 2.499804 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:49 INFO: Evaluating on dev set...
-2026-01-16 07:01:49 INFO: LAS	MLAS	BLEX
-2026-01-16 07:01:49 INFO: 12.87	8.68	11.44
-2026-01-16 07:01:49 INFO: step 3800: train_loss = 2.251492, dev_score = 0.1287
-2026-01-16 07:01:50 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:01:50 INFO: new model checkpoint saved.
-2026-01-16 07:01:50 INFO: Finished STEP 3820/50000, loss = 4.027611 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:51 INFO: Finished STEP 3840/50000, loss = 0.331204 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:01:52 INFO: Finished STEP 3860/50000, loss = 1.255922 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:01:52 INFO: Finished STEP 3880/50000, loss = 1.898022 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:53 INFO: Finished STEP 3900/50000, loss = 1.984990 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:01:53 INFO: Evaluating on dev set...
-2026-01-16 07:01:53 INFO: LAS	MLAS	BLEX
-2026-01-16 07:01:53 INFO: 14.36	8.71	9.90
-2026-01-16 07:01:53 INFO: step 3900: train_loss = 2.995180, dev_score = 0.1436
-2026-01-16 07:01:54 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:01:54 INFO: new model checkpoint saved.
-2026-01-16 07:01:54 INFO: Finished STEP 3920/50000, loss = 2.342984 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:55 INFO: Finished STEP 3940/50000, loss = 0.587340 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:56 INFO: Finished STEP 3960/50000, loss = 1.429672 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:01:56 INFO: Finished STEP 3980/50000, loss = 1.932409 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:01:57 INFO: Finished STEP 4000/50000, loss = 2.074286 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:01:57 INFO: Evaluating on dev set...
-2026-01-16 07:01:57 INFO: LAS	MLAS	BLEX
-2026-01-16 07:01:57 INFO: 13.61	7.11	9.49
-2026-01-16 07:01:57 INFO: step 4000: train_loss = 2.342431, dev_score = 0.1361
-2026-01-16 07:01:58 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:01:58 INFO: new model checkpoint saved.
-2026-01-16 07:01:58 INFO: Finished STEP 4020/50000, loss = 0.798264 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:01:59 INFO: Finished STEP 4040/50000, loss = 0.128710 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:02:00 INFO: Finished STEP 4060/50000, loss = 0.836676 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:02:00 INFO: Finished STEP 4080/50000, loss = 1.385599 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:02:01 INFO: Finished STEP 4100/50000, loss = 1.616785 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:02:01 INFO: Evaluating on dev set...
-2026-01-16 07:02:01 INFO: LAS	MLAS	BLEX
-2026-01-16 07:02:01 INFO: 17.33	6.88	8.91
-2026-01-16 07:02:01 INFO: step 4100: train_loss = 2.344978, dev_score = 0.1733
-2026-01-16 07:02:01 INFO: Training ended with 4100 steps.
-2026-01-16 07:02:01 INFO: Best dev F1 = 23.51, at iteration = 3100
-2026-01-16 07:02:02 INFO: Running dev depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '8', '--dropout', '0.33']
-2026-01-16 07:02:02 INFO: Running parser in predict mode
-2026-01-16 07:02:02 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:02:04 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2026-01-16 07:02:04 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:02:04 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2026-01-16 07:02:04 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:02:04 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2026-01-16 07:02:04 INFO: Loading data with batch size 8...
-2026-01-16 07:02:04 DEBUG: 9 batches created.
-2026-01-16 07:02:05 INFO: F1 scores for each dependency:
-  Note that unlabeled attachment errors hurt the labeled attachment scores
-       acl: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
- acl:relcl: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
-     advcl: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
-    advmod: p 0.1000 r 0.1200 f1 0.1091 (25 actual)
-      amod: p 0.0000 r 0.0000 f1 0.0000 (31 actual)
-     appos: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-       aux: p 0.0000 r 0.0000 f1 0.0000 (11 actual)
-      case: p 0.4557 r 0.6429 f1 0.5333 (56 actual)
-        cc: p 0.0000 r 0.0000 f1 0.0000 (13 actual)
-     ccomp: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-      conj: p 0.0000 r 0.0000 f1 0.0000 (12 actual)
-       cop: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
-     csubj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-       det: p 0.5000 r 0.4091 f1 0.4500 (22 actual)
-      expl: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
- flat:name: p 0.0000 r 0.0000 f1 0.0000 (0 actual)
-      iobj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-      mark: p 0.0000 r 0.0000 f1 0.0000 (12 actual)
-      nmod: p 0.0000 r 0.0000 f1 0.0000 (15 actual)
- nmod:poss: p 0.0000 r 0.0000 f1 0.0000 (19 actual)
-     nsubj: p 0.0641 r 0.5882 f1 0.1156 (17 actual)
-nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
-       obj: p 0.0000 r 0.0000 f1 0.0000 (22 actual)
-       obl: p 0.2373 r 0.3415 f1 0.2800 (41 actual)
- obl:agent: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-    orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
- parataxis: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
-     punct: p 0.3077 r 0.3077 f1 0.3077 (52 actual)
-      root: p 0.5556 r 0.5556 f1 0.5556 (9 actual)
-     xcomp: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
-2026-01-16 07:02:05 INFO: LAS	MLAS	BLEX
-2026-01-16 07:02:05 INFO: 23.51	9.39	13.88
-2026-01-16 07:02:05 INFO: Parser score:
-2026-01-16 07:02:05 INFO: sv_diachronic 23.51
-2026-01-16 07:02:05 INFO: Finished running dev set on
-UD_Swedish-diachronic
-  UAS   LAS  CLAS  MLAS  BLEX
-48.27 23.51 13.88  9.39 13.88
-2026-01-16 07:02:05 INFO: Running test depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '8', '--dropout', '0.33']
-2026-01-16 07:02:05 INFO: Running parser in predict mode
-2026-01-16 07:02:05 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:02:07 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2026-01-16 07:02:07 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:02:07 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2026-01-16 07:02:07 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:02:07 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2026-01-16 07:02:07 INFO: Loading data with batch size 8...
-2026-01-16 07:02:07 DEBUG: 99 batches created.
-2026-01-16 07:02:11 INFO: F1 scores for each dependency:
-  Note that unlabeled attachment errors hurt the labeled attachment scores
-         acl: p 0.0000 r 0.0000 f1 0.0000 (32 actual)
-   acl:cleft: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-   acl:relcl: p 0.0000 r 0.0000 f1 0.0000 (75 actual)
-       advcl: p 0.0000 r 0.0000 f1 0.0000 (60 actual)
- advcl:relcl: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-      advmod: p 0.0662 r 0.1007 f1 0.0799 (268 actual)
-        amod: p 0.0000 r 0.0000 f1 0.0000 (230 actual)
-       appos: p 0.0000 r 0.0000 f1 0.0000 (13 actual)
-         aux: p 0.0000 r 0.0000 f1 0.0000 (84 actual)
-    aux:pass: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-        case: p 0.3323 r 0.5818 f1 0.4230 (373 actual)
-          cc: p 0.0000 r 0.0000 f1 0.0000 (155 actual)
-       ccomp: p 0.0000 r 0.0000 f1 0.0000 (35 actual)
-compound:prt: p 0.0000 r 0.0000 f1 0.0000 (21 actual)
-        conj: p 0.0000 r 0.0000 f1 0.0000 (158 actual)
-         cop: p 0.0000 r 0.0000 f1 0.0000 (46 actual)
-       csubj: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-         dep: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-         det: p 0.2518 r 0.1683 f1 0.2017 (208 actual)
-   discourse: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
-  dislocated: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-        expl: p 0.0000 r 0.0000 f1 0.0000 (11 actual)
-     expl:pv: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-       fixed: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
-        flat: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-   flat:name: p 0.4000 r 0.1667 f1 0.2353 (12 actual)
-    goeswith: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-        iobj: p 0.0000 r 0.0000 f1 0.0000 (14 actual)
-        mark: p 0.0000 r 0.0000 f1 0.0000 (153 actual)
-        nmod: p 0.0000 r 0.0000 f1 0.0000 (102 actual)
-   nmod:poss: p 0.0000 r 0.0000 f1 0.0000 (142 actual)
-       nsubj: p 0.1027 r 0.5464 f1 0.1729 (280 actual)
-  nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (25 actual)
-      nummod: p 0.0000 r 0.0000 f1 0.0000 (10 actual)
-         obj: p 0.0000 r 0.0000 f1 0.0000 (183 actual)
-         obl: p 0.2272 r 0.3309 f1 0.2694 (278 actual)
-   obl:agent: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-      orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-   parataxis: p 0.0000 r 0.0000 f1 0.0000 (18 actual)
-       punct: p 0.3411 r 0.3459 f1 0.3435 (425 actual)
-  reparandum: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-        root: p 0.4545 r 0.4545 f1 0.4545 (99 actual)
-    vocative: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
-       xcomp: p 0.0000 r 0.0000 f1 0.0000 (75 actual)
-2026-01-16 07:02:11 INFO: LAS	MLAS	BLEX
-2026-01-16 07:02:11 INFO: 20.06	11.15	14.33
-2026-01-16 07:02:11 INFO: Parser score:
-2026-01-16 07:02:11 INFO: sv_diachronic 20.06
-2026-01-16 07:02:11 INFO: Finished running test set on
-UD_Swedish-diachronic
-  UAS   LAS  CLAS  MLAS  BLEX
-48.02 20.06 14.33 11.15 14.33
-DONE.
-Full log saved to: logs/log_conll17.pt_sv_diachron_20260116_065352.txt
-Symlink updated: logs/latest.txt → log_conll17.pt_sv_diachron_20260116_065352.txt

logs/log_conll17.pt_sv_diachron_20260116_073035.txt

@@ -1,868 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_diachron_20260116_073035.txt ===
-Language codes: sv diachron
-Using pretrained model: conll17.pt
-
-Running: python prepare-train-val-test.py sv diachron
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec991-spf148.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec988-spf145.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec1102-spf259.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-letter141673-Stalhammar.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec1033-spf190.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec25-Runius.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec987-spf144.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec1063-spf220.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Including DigPhil MACHINE in TRAIN (minus gold)…
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
-Cleaning TRAIN...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-[REMOVED] sent_id=6  ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
-[REMOVED] sent_id=7_8  ERRORS=['Multiple roots found: [5, 10]']
-[REMOVED] sent_id=30_31  ERRORS=['Multiple roots found: [3, 18]']
-[REMOVED] sent_id=35  ERRORS=['Line 36: Invalid token ID or head']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [7, 20]']
-[REMOVED] sent_id=8_9  ERRORS=['Multiple roots found: [24, 57]']
-[REMOVED] sent_id=12_13  ERRORS=['Multiple roots found: [11, 16]']
-[REMOVED] sent_id=124_split2  ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
-[REMOVED] sent_id=396  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=416  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=589  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=909  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=912  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=3_split1  ERRORS=['Multiple roots found: [4, 15, 17]']
-[REMOVED] sent_id=3_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
-[REMOVED] sent_id=3_4  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=5_6  ERRORS=['Multiple roots found: [3, 24]']
-[REMOVED] sent_id=11_12_13  ERRORS=['Multiple roots found: [5, 17, 25]']
-[REMOVED] sent_id=119  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=188  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=223  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=268  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=325  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=388  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=399  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=475  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=505  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=520  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=562  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=711  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=731  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=867  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=884  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=923  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=939  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1086  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1251  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1345  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1459  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1656  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=87_88  ERRORS=['Multiple roots found: [3, 6]']
-[REMOVED] sent_id=65_split2_66_split2  ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
-[REMOVED] sent_id=25  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=136  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=208  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=230  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=245  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=276  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=320  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=366  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=519  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=569  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=50_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
-[REMOVED] sent_id=53_54  ERRORS=['Multiple roots found: [27, 91]']
-[REMOVED] sent_id=55_56_57  ERRORS=['Multiple roots found: [2, 4, 13]']
-[REMOVED] sent_id=17_split1  ERRORS=['Multiple roots found: [2, 14, 17]']
-[REMOVED] sent_id=17_split2  ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
-[REMOVED] sent_id=19_split1  ERRORS=['Multiple roots found: [3, 31]']
-Cleaning DEV...
-Cleaning TEST...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Writing TRAIN → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (60334 valid sentences)
-Writing DEV → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (34 valid sentences)
-Writing TEST → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (74 valid sentences)
-Done.
-Sourcing scripts/config_alvis.sh
-Running stanza dataset preparation…
-2026-01-16 07:30:43 INFO: Datasets program called with:
-/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2026-01-16 07:30:43 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
-
-2026-01-16 07:30:44 INFO: Downloaded file to /cephyr/users/cleland/Alvis/stanza_resources/resources.json
-2026-01-16 07:30:44 DEBUG: Processing parameter "processors"...
-2026-01-16 07:30:44 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
-2026-01-16 07:30:44 INFO: Downloading these customized packages for language: sv (Swedish)...
-=======================
-| Processor | Package |
------------------------
-=======================
-
-2026-01-16 07:30:44 INFO: Finished downloading models and saved to /cephyr/users/cleland/Alvis/stanza_resources
-2026-01-16 07:30:44 INFO: Using tagger model in /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt for sv_diachronic
-2026-01-16 07:30:44 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2026-01-16 07:30:44 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-Augmented 189 quotes: Counter({'""': 26, '「」': 23, '„”': 23, '»«': 22, '″″': 18, '””': 18, '““': 17, '《》': 17, '„“': 13, '«»': 12})
-2026-01-16 07:30:46 INFO: Running tagger to retag /local/tmp.5658622/tmp0y7bjycl/sv_diachronic.train.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5658622/tmp0y7bjycl/sv_diachronic.train.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu']
-2026-01-16 07:30:47 INFO: Running tagger in predict mode
-2026-01-16 07:30:47 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2026-01-16 07:30:48 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2026-01-16 07:30:48 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:30:48 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2026-01-16 07:30:48 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:30:49 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2026-01-16 07:30:51 INFO: Loading data with batch size 250...
-2026-01-16 07:31:23 INFO: Start evaluation...
-2026-01-16 07:34:24 INFO: UPOS	XPOS	UFeats	AllTags
-2026-01-16 07:34:24 INFO: 99.37	87.38	98.56	87.09
-2026-01-16 07:34:24 INFO: POS Tagger score: sv_diachronic 87.09
-2026-01-16 07:34:25 INFO: Running tagger to retag /local/tmp.5658622/tmp0y7bjycl/sv_diachronic.dev.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5658622/tmp0y7bjycl/sv_diachronic.dev.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu']
-2026-01-16 07:34:25 INFO: Running tagger in predict mode
-2026-01-16 07:34:25 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2026-01-16 07:34:26 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2026-01-16 07:34:26 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:34:26 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2026-01-16 07:34:26 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:34:26 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2026-01-16 07:34:26 INFO: Loading data with batch size 250...
-2026-01-16 07:34:26 INFO: Start evaluation...
-2026-01-16 07:34:27 INFO: UPOS	XPOS	UFeats	AllTags
-2026-01-16 07:34:27 INFO: 90.62	97.55	94.98	89.40
-2026-01-16 07:34:27 INFO: POS Tagger score: sv_diachronic 89.40
-2026-01-16 07:34:27 INFO: Running tagger to retag /local/tmp.5658622/tmp0y7bjycl/sv_diachronic.test.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5658622/tmp0y7bjycl/sv_diachronic.test.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu']
-2026-01-16 07:34:27 INFO: Running tagger in predict mode
-2026-01-16 07:34:27 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2026-01-16 07:34:29 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2026-01-16 07:34:29 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:34:29 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2026-01-16 07:34:29 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:34:29 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2026-01-16 07:34:29 INFO: Loading data with batch size 250...
-2026-01-16 07:34:29 INFO: Start evaluation...
-2026-01-16 07:34:29 INFO: UPOS	XPOS	UFeats	AllTags
-2026-01-16 07:34:29 INFO: 94.88	95.25	95.21	90.09
-2026-01-16 07:34:29 INFO: POS Tagger score: sv_diachronic 90.09
-Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /local/tmp.5658622/tmp0y7bjycl/sv_diachronic.train.gold.conllu
-Swapped 'w1, w2' for 'w1 ,w2' 106 times
-Added 520 new sentences with asdf, zzzz -> asdf,zzzz
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /local/tmp.5658622/tmp0y7bjycl/sv_diachronic.dev.gold.conllu
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /local/tmp.5658622/tmp0y7bjycl/sv_diachronic.test.gold.conllu
-Running stanza dependency parser training…
-2026-01-16 07:34:42 INFO: Training program called with:
-/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt --batch_size 8 --dropout 0.15
-2026-01-16 07:34:42 DEBUG: UD_Swedish-diachronic: sv_diachronic
-2026-01-16 07:34:42 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2026-01-16 07:34:42 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2026-01-16 07:34:42 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_charlm_parser.pt does not exist, training new model
-2026-01-16 07:34:42 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2026-01-16 07:34:42 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2026-01-16 07:34:42 INFO: Running train depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--train_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--batch_size', '5000', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'train', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '8', '--dropout', '0.15']
-2026-01-16 07:34:42 INFO: Running parser in train mode
-2026-01-16 07:34:42 INFO: Using pretrained contextualized char embedding
-2026-01-16 07:34:42 INFO: Loading data with batch size 8...
-2026-01-16 07:34:48 INFO: Train File /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu, Data Size: 60854
-2026-01-16 07:34:48 INFO: Original data size: 60854
-2026-01-16 07:34:48 INFO: Augmented data size: 60854
-2026-01-16 07:35:03 WARNING: sv_diachronic is not a known dataset.  Examining the data to choose which xpos vocab to use
-2026-01-16 07:35:03 INFO: Original length = 60854
-2026-01-16 07:35:03 INFO: Filtered length = 60854
-2026-01-16 07:35:17 WARNING: Chose XPOSDescription(xpos_type=<XPOSType.XPOS: 1>, sep='|') for the xpos factory for sv_diachronic
-2026-01-16 07:35:22 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2026-01-16 07:35:33 DEBUG: 58067 batches created.
-2026-01-16 07:35:33 DEBUG: 34 batches created.
-2026-01-16 07:35:33 INFO: Training parser...
-2026-01-16 07:35:33 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:35:33 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2026-01-16 07:35:34 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:35:34 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2026-01-16 07:35:37 INFO: Finished STEP 20/50000, loss = 0.000000 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:35:38 INFO: Finished STEP 40/50000, loss = 0.000000 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:35:38 INFO: Finished STEP 60/50000, loss = 0.000000 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:35:39 INFO: Finished STEP 80/50000, loss = 0.000000 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:35:39 INFO: Finished STEP 100/50000, loss = 0.000000 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:35:39 INFO: Evaluating on dev set...
-2026-01-16 07:35:41 INFO: LAS	MLAS	BLEX
-2026-01-16 07:35:41 INFO: 0.06	0.08	0.08
-2026-01-16 07:35:41 INFO: step 100: train_loss = 379.523051, dev_score = 0.0006
-2026-01-16 07:35:41 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:35:41 INFO: new best model saved.
-2026-01-16 07:35:42 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:35:42 INFO: new model checkpoint saved.
-2026-01-16 07:35:42 INFO: Finished STEP 120/50000, loss = 0.000000 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:35:43 INFO: Finished STEP 140/50000, loss = 0.000000 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:35:44 INFO: Finished STEP 160/50000, loss = 0.000000 (0.028 sec/batch), lr: 0.003000
-2026-01-16 07:35:44 INFO: Finished STEP 180/50000, loss = 0.000000 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:35:45 INFO: Finished STEP 200/50000, loss = 0.000000 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:35:45 INFO: Evaluating on dev set...
-2026-01-16 07:35:46 INFO: LAS	MLAS	BLEX
-2026-01-16 07:35:46 INFO: 0.06	0.08	0.08
-2026-01-16 07:35:46 INFO: step 200: train_loss = 0.000000, dev_score = 0.0006
-2026-01-16 07:35:47 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:35:47 INFO: new best model saved.
-2026-01-16 07:35:47 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:35:47 INFO: new model checkpoint saved.
-2026-01-16 07:35:48 INFO: Finished STEP 220/50000, loss = 0.000000 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:35:48 INFO: Finished STEP 240/50000, loss = 0.000000 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:35:49 INFO: Finished STEP 260/50000, loss = 0.000000 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:35:50 INFO: Finished STEP 280/50000, loss = 0.000000 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:35:50 INFO: Finished STEP 300/50000, loss = 2.424209 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:35:50 INFO: Evaluating on dev set...
-2026-01-16 07:35:52 INFO: LAS	MLAS	BLEX
-2026-01-16 07:35:52 INFO: 1.35	0.00	0.00
-2026-01-16 07:35:52 INFO: step 300: train_loss = 158.649473, dev_score = 0.0135
-2026-01-16 07:35:52 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:35:52 INFO: new best model saved.
-2026-01-16 07:35:53 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:35:53 INFO: new model checkpoint saved.
-2026-01-16 07:35:53 INFO: Finished STEP 320/50000, loss = 1.969488 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:35:54 INFO: Finished STEP 340/50000, loss = 1.649248 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:35:54 INFO: Finished STEP 360/50000, loss = 1.934138 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:35:55 INFO: Finished STEP 380/50000, loss = 2.360754 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:35:56 INFO: Finished STEP 400/50000, loss = 0.562345 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:35:56 INFO: Evaluating on dev set...
-2026-01-16 07:35:58 INFO: LAS	MLAS	BLEX
-2026-01-16 07:35:58 INFO: 4.72	4.16	4.25
-2026-01-16 07:35:58 INFO: step 400: train_loss = 2.121128, dev_score = 0.0472
-2026-01-16 07:35:58 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:35:58 INFO: new best model saved.
-2026-01-16 07:35:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:35:59 INFO: new model checkpoint saved.
-2026-01-16 07:35:59 INFO: Finished STEP 420/50000, loss = 0.002566 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:00 INFO: Finished STEP 440/50000, loss = 0.734853 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:01 INFO: Finished STEP 460/50000, loss = 0.494820 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:01 INFO: Finished STEP 480/50000, loss = 3.201444 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:02 INFO: Finished STEP 500/50000, loss = 0.795342 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:02 INFO: Evaluating on dev set...
-2026-01-16 07:36:04 INFO: LAS	MLAS	BLEX
-2026-01-16 07:36:04 INFO: 5.88	5.61	6.11
-2026-01-16 07:36:04 INFO: step 500: train_loss = 1.037986, dev_score = 0.0588
-2026-01-16 07:36:04 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:36:04 INFO: new best model saved.
-2026-01-16 07:36:05 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:36:05 INFO: new model checkpoint saved.
-2026-01-16 07:36:05 INFO: Finished STEP 520/50000, loss = 0.778292 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:06 INFO: Finished STEP 540/50000, loss = 0.409450 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:06 INFO: Finished STEP 560/50000, loss = 0.671611 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:07 INFO: Finished STEP 580/50000, loss = 0.081582 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:08 INFO: Finished STEP 600/50000, loss = 0.436447 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:08 INFO: Evaluating on dev set...
-2026-01-16 07:36:10 INFO: LAS	MLAS	BLEX
-2026-01-16 07:36:10 INFO: 5.76	6.09	6.18
-2026-01-16 07:36:10 INFO: step 600: train_loss = 1.044479, dev_score = 0.0576
-2026-01-16 07:36:11 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:36:11 INFO: new model checkpoint saved.
-2026-01-16 07:36:11 INFO: Finished STEP 620/50000, loss = 0.887764 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:36:12 INFO: Finished STEP 640/50000, loss = 0.527841 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:12 INFO: Finished STEP 660/50000, loss = 0.269212 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:13 INFO: Finished STEP 680/50000, loss = 2.426815 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:14 INFO: Finished STEP 700/50000, loss = 0.000517 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:14 INFO: Evaluating on dev set...
-2026-01-16 07:36:15 INFO: LAS	MLAS	BLEX
-2026-01-16 07:36:15 INFO: 5.21	4.95	5.37
-2026-01-16 07:36:15 INFO: step 700: train_loss = 0.978077, dev_score = 0.0521
-2026-01-16 07:36:16 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:36:16 INFO: new model checkpoint saved.
-2026-01-16 07:36:16 INFO: Finished STEP 720/50000, loss = 2.358844 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:17 INFO: Finished STEP 740/50000, loss = 0.791893 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:18 INFO: Finished STEP 760/50000, loss = 0.906499 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:18 INFO: Finished STEP 780/50000, loss = 1.075008 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:19 INFO: Finished STEP 800/50000, loss = 0.000000 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:19 INFO: Evaluating on dev set...
-2026-01-16 07:36:21 INFO: LAS	MLAS	BLEX
-2026-01-16 07:36:21 INFO: 2.14	1.13	1.13
-2026-01-16 07:36:21 INFO: step 800: train_loss = 0.975283, dev_score = 0.0214
-2026-01-16 07:36:21 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:36:21 INFO: new model checkpoint saved.
-2026-01-16 07:36:22 INFO: Finished STEP 820/50000, loss = 0.678568 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:22 INFO: Finished STEP 840/50000, loss = 0.000088 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:23 INFO: Finished STEP 860/50000, loss = 1.472390 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:24 INFO: Finished STEP 880/50000, loss = 0.818992 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:24 INFO: Finished STEP 900/50000, loss = 1.486508 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:36:24 INFO: Evaluating on dev set...
-2026-01-16 07:36:26 INFO: LAS	MLAS	BLEX
-2026-01-16 07:36:26 INFO: 3.98	1.89	3.04
-2026-01-16 07:36:26 INFO: step 900: train_loss = 0.911072, dev_score = 0.0398
-2026-01-16 07:36:27 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:36:27 INFO: new model checkpoint saved.
-2026-01-16 07:36:27 INFO: Finished STEP 920/50000, loss = 0.001453 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:28 INFO: Finished STEP 940/50000, loss = 0.934357 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:28 INFO: Finished STEP 960/50000, loss = 0.010020 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:29 INFO: Finished STEP 980/50000, loss = 0.627853 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:30 INFO: Finished STEP 1000/50000, loss = 0.240197 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:30 INFO: Evaluating on dev set...
-2026-01-16 07:36:31 INFO: LAS	MLAS	BLEX
-2026-01-16 07:36:31 INFO: 2.76	1.78	1.78
-2026-01-16 07:36:31 INFO: step 1000: train_loss = 0.752214, dev_score = 0.0276
-2026-01-16 07:36:32 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:36:32 INFO: new model checkpoint saved.
-2026-01-16 07:36:32 INFO: Finished STEP 1020/50000, loss = 1.534254 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:33 INFO: Finished STEP 1040/50000, loss = 0.000038 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:34 INFO: Finished STEP 1060/50000, loss = 1.659585 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:34 INFO: Finished STEP 1080/50000, loss = 0.403725 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:35 INFO: Finished STEP 1100/50000, loss = 0.000000 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:36:35 INFO: Evaluating on dev set...
-2026-01-16 07:36:37 INFO: LAS	MLAS	BLEX
-2026-01-16 07:36:37 INFO: 3.12	2.80	2.89
-2026-01-16 07:36:37 INFO: step 1100: train_loss = 0.824820, dev_score = 0.0312
-2026-01-16 07:36:37 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:36:37 INFO: new model checkpoint saved.
-2026-01-16 07:36:38 INFO: Finished STEP 1120/50000, loss = 6.452912 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:36:38 INFO: Finished STEP 1140/50000, loss = 0.749853 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:36:39 INFO: Finished STEP 1160/50000, loss = 4.511549 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:40 INFO: Finished STEP 1180/50000, loss = 2.717284 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:40 INFO: Finished STEP 1200/50000, loss = 0.523814 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:36:40 INFO: Evaluating on dev set...
-2026-01-16 07:36:42 INFO: LAS	MLAS	BLEX
-2026-01-16 07:36:42 INFO: 4.53	1.41	1.50
-2026-01-16 07:36:42 INFO: step 1200: train_loss = 2.303929, dev_score = 0.0453
-2026-01-16 07:36:42 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:36:42 INFO: new model checkpoint saved.
-2026-01-16 07:36:43 INFO: Finished STEP 1220/50000, loss = 0.628675 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:44 INFO: Finished STEP 1240/50000, loss = 1.167853 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:36:44 INFO: Finished STEP 1260/50000, loss = 2.458635 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:36:45 INFO: Finished STEP 1280/50000, loss = 1.727715 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:46 INFO: Finished STEP 1300/50000, loss = 3.715376 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:36:46 INFO: Evaluating on dev set...
-2026-01-16 07:36:47 INFO: LAS	MLAS	BLEX
-2026-01-16 07:36:47 INFO: 3.06	1.24	1.32
-2026-01-16 07:36:47 INFO: step 1300: train_loss = 1.477564, dev_score = 0.0306
-2026-01-16 07:36:48 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:36:48 INFO: new model checkpoint saved.
-2026-01-16 07:36:48 INFO: Finished STEP 1320/50000, loss = 1.768359 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:49 INFO: Finished STEP 1340/50000, loss = 1.245610 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:50 INFO: Finished STEP 1360/50000, loss = 0.451835 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:36:50 INFO: Finished STEP 1380/50000, loss = 1.443755 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:36:51 INFO: Finished STEP 1400/50000, loss = 1.238335 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:51 INFO: Evaluating on dev set...
-2026-01-16 07:36:53 INFO: LAS	MLAS	BLEX
-2026-01-16 07:36:53 INFO: 7.11	6.09	6.96
-2026-01-16 07:36:53 INFO: step 1400: train_loss = 1.361486, dev_score = 0.0711
-2026-01-16 07:36:53 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:36:53 INFO: new best model saved.
-2026-01-16 07:36:53 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:36:53 INFO: new model checkpoint saved.
-2026-01-16 07:36:54 INFO: Finished STEP 1420/50000, loss = 1.591924 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:36:55 INFO: Finished STEP 1440/50000, loss = 0.735520 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:36:55 INFO: Finished STEP 1460/50000, loss = 1.212776 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:36:56 INFO: Finished STEP 1480/50000, loss = 1.210722 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:36:57 INFO: Finished STEP 1500/50000, loss = 3.332783 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:36:57 INFO: Evaluating on dev set...
-2026-01-16 07:36:58 INFO: LAS	MLAS	BLEX
-2026-01-16 07:36:58 INFO: 6.74	3.03	3.88
-2026-01-16 07:36:58 INFO: step 1500: train_loss = 1.381426, dev_score = 0.0674
-2026-01-16 07:36:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:36:59 INFO: new model checkpoint saved.
-2026-01-16 07:37:00 INFO: Finished STEP 1520/50000, loss = 0.796412 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:00 INFO: Finished STEP 1540/50000, loss = 0.730972 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:01 INFO: Finished STEP 1560/50000, loss = 1.262556 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:01 INFO: Finished STEP 1580/50000, loss = 0.734725 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:02 INFO: Finished STEP 1600/50000, loss = 0.073490 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:02 INFO: Evaluating on dev set...
-2026-01-16 07:37:05 INFO: LAS	MLAS	BLEX
-2026-01-16 07:37:05 INFO: 12.87	10.38	11.31
-2026-01-16 07:37:05 INFO: step 1600: train_loss = 1.284919, dev_score = 0.1287
-2026-01-16 07:37:06 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:37:06 INFO: new best model saved.
-2026-01-16 07:37:06 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:37:06 INFO: new model checkpoint saved.
-2026-01-16 07:37:07 INFO: Finished STEP 1620/50000, loss = 0.640151 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:07 INFO: Finished STEP 1640/50000, loss = 1.584237 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:08 INFO: Finished STEP 1660/50000, loss = 2.824079 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:09 INFO: Finished STEP 1680/50000, loss = 1.456016 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:09 INFO: Finished STEP 1700/50000, loss = 1.605577 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:09 INFO: Evaluating on dev set...
-2026-01-16 07:37:11 INFO: LAS	MLAS	BLEX
-2026-01-16 07:37:11 INFO: 13.36	9.21	10.85
-2026-01-16 07:37:11 INFO: step 1700: train_loss = 1.177195, dev_score = 0.1336
-2026-01-16 07:37:11 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:37:11 INFO: new best model saved.
-2026-01-16 07:37:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:37:12 INFO: new model checkpoint saved.
-2026-01-16 07:37:12 INFO: Finished STEP 1720/50000, loss = 1.005753 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:13 INFO: Finished STEP 1740/50000, loss = 1.587673 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:37:14 INFO: Finished STEP 1760/50000, loss = 1.467617 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:14 INFO: Finished STEP 1780/50000, loss = 2.158596 (0.034 sec/batch), lr: 0.003000
-2026-01-16 07:37:15 INFO: Finished STEP 1800/50000, loss = 0.786507 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:15 INFO: Evaluating on dev set...
-2026-01-16 07:37:17 INFO: LAS	MLAS	BLEX
-2026-01-16 07:37:17 INFO: 10.11	6.65	7.70
-2026-01-16 07:37:17 INFO: step 1800: train_loss = 1.175222, dev_score = 0.1011
-2026-01-16 07:37:17 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:37:17 INFO: new model checkpoint saved.
-2026-01-16 07:37:18 INFO: Finished STEP 1820/50000, loss = 0.861782 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:19 INFO: Finished STEP 1840/50000, loss = 1.011274 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:19 INFO: Finished STEP 1860/50000, loss = 0.641151 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:20 INFO: Finished STEP 1880/50000, loss = 0.764767 (0.036 sec/batch), lr: 0.003000
-2026-01-16 07:37:21 INFO: Finished STEP 1900/50000, loss = 1.876217 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:21 INFO: Evaluating on dev set...
-2026-01-16 07:37:25 INFO: LAS	MLAS	BLEX
-2026-01-16 07:37:25 INFO: 7.66	4.98	5.64
-2026-01-16 07:37:25 INFO: step 1900: train_loss = 1.227179, dev_score = 0.0766
-2026-01-16 07:37:26 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:37:26 INFO: new model checkpoint saved.
-2026-01-16 07:37:27 INFO: Finished STEP 1920/50000, loss = 1.518084 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:27 INFO: Finished STEP 1940/50000, loss = 0.689516 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:28 INFO: Finished STEP 1960/50000, loss = 0.541874 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:29 INFO: Finished STEP 1980/50000, loss = 1.434824 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:29 INFO: Finished STEP 2000/50000, loss = 0.343702 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:29 INFO: Evaluating on dev set...
-2026-01-16 07:37:31 INFO: LAS	MLAS	BLEX
-2026-01-16 07:37:31 INFO: 11.52	4.14	5.58
-2026-01-16 07:37:31 INFO: step 2000: train_loss = 1.288903, dev_score = 0.1152
-2026-01-16 07:37:32 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:37:32 INFO: new model checkpoint saved.
-2026-01-16 07:37:33 INFO: Finished STEP 2020/50000, loss = 1.551595 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:33 INFO: Finished STEP 2040/50000, loss = 0.806195 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:34 INFO: Finished STEP 2060/50000, loss = 0.586268 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:34 INFO: Finished STEP 2080/50000, loss = 1.326350 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:35 INFO: Finished STEP 2100/50000, loss = 1.068810 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:35 INFO: Evaluating on dev set...
-2026-01-16 07:37:37 INFO: LAS	MLAS	BLEX
-2026-01-16 07:37:37 INFO: 3.49	1.97	2.55
-2026-01-16 07:37:37 INFO: step 2100: train_loss = 1.064336, dev_score = 0.0349
-2026-01-16 07:37:37 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:37:37 INFO: new model checkpoint saved.
-2026-01-16 07:37:38 INFO: Finished STEP 2120/50000, loss = 0.723737 (0.033 sec/batch), lr: 0.003000
-2026-01-16 07:37:39 INFO: Finished STEP 2140/50000, loss = 1.762422 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:39 INFO: Finished STEP 2160/50000, loss = 0.039095 (0.036 sec/batch), lr: 0.003000
-2026-01-16 07:37:40 INFO: Finished STEP 2180/50000, loss = 0.746787 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:41 INFO: Finished STEP 2200/50000, loss = 0.145349 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:41 INFO: Evaluating on dev set...
-2026-01-16 07:37:43 INFO: LAS	MLAS	BLEX
-2026-01-16 07:37:43 INFO: 13.30	9.26	9.89
-2026-01-16 07:37:43 INFO: step 2200: train_loss = 1.202942, dev_score = 0.1330
-2026-01-16 07:37:43 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:37:43 INFO: new model checkpoint saved.
-2026-01-16 07:37:44 INFO: Finished STEP 2220/50000, loss = 2.421659 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:44 INFO: Finished STEP 2240/50000, loss = 2.592595 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:45 INFO: Finished STEP 2260/50000, loss = 3.920859 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:46 INFO: Finished STEP 2280/50000, loss = 1.928826 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:46 INFO: Finished STEP 2300/50000, loss = 5.176385 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:46 INFO: Evaluating on dev set...
-2026-01-16 07:37:48 INFO: LAS	MLAS	BLEX
-2026-01-16 07:37:48 INFO: 17.83	9.07	10.24
-2026-01-16 07:37:48 INFO: step 2300: train_loss = 2.391255, dev_score = 0.1783
-2026-01-16 07:37:48 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:37:48 INFO: new best model saved.
-2026-01-16 07:37:49 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:37:49 INFO: new model checkpoint saved.
-2026-01-16 07:37:49 INFO: Finished STEP 2320/50000, loss = 2.267345 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:50 INFO: Finished STEP 2340/50000, loss = 0.719973 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:51 INFO: Finished STEP 2360/50000, loss = 1.036772 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:51 INFO: Finished STEP 2380/50000, loss = 2.774978 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:52 INFO: Finished STEP 2400/50000, loss = 1.541276 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:52 INFO: Evaluating on dev set...
-2026-01-16 07:37:54 INFO: LAS	MLAS	BLEX
-2026-01-16 07:37:54 INFO: 5.76	3.46	4.14
-2026-01-16 07:37:54 INFO: step 2400: train_loss = 2.107856, dev_score = 0.0576
-2026-01-16 07:37:54 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:37:54 INFO: new model checkpoint saved.
-2026-01-16 07:37:55 INFO: Finished STEP 2420/50000, loss = 0.558537 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:37:55 INFO: Finished STEP 2440/50000, loss = 0.417549 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:37:56 INFO: Finished STEP 2460/50000, loss = 1.211123 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:37:57 INFO: Finished STEP 2480/50000, loss = 1.066298 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:37:57 INFO: Finished STEP 2500/50000, loss = 1.788547 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:37:57 INFO: Evaluating on dev set...
-2026-01-16 07:37:59 INFO: LAS	MLAS	BLEX
-2026-01-16 07:37:59 INFO: 12.68	5.16	6.11
-2026-01-16 07:37:59 INFO: step 2500: train_loss = 2.112291, dev_score = 0.1268
-2026-01-16 07:38:00 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:38:00 INFO: new model checkpoint saved.
-2026-01-16 07:38:00 INFO: Finished STEP 2520/50000, loss = 1.717021 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:38:01 INFO: Finished STEP 2540/50000, loss = 1.660125 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:38:01 INFO: Finished STEP 2560/50000, loss = 1.768633 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:02 INFO: Finished STEP 2580/50000, loss = 0.593540 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:38:03 INFO: Finished STEP 2600/50000, loss = 0.938626 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:03 INFO: Evaluating on dev set...
-2026-01-16 07:38:05 INFO: LAS	MLAS	BLEX
-2026-01-16 07:38:05 INFO: 19.12	9.02	10.39
-2026-01-16 07:38:05 INFO: step 2600: train_loss = 1.933433, dev_score = 0.1912
-2026-01-16 07:38:05 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:38:05 INFO: new best model saved.
-2026-01-16 07:38:06 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:38:06 INFO: new model checkpoint saved.
-2026-01-16 07:38:06 INFO: Finished STEP 2620/50000, loss = 2.193763 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:38:07 INFO: Finished STEP 2640/50000, loss = 1.114236 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:07 INFO: Finished STEP 2660/50000, loss = 1.077905 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:08 INFO: Finished STEP 2680/50000, loss = 2.145391 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:38:09 INFO: Finished STEP 2700/50000, loss = 4.426371 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:09 INFO: Evaluating on dev set...
-2026-01-16 07:38:11 INFO: LAS	MLAS	BLEX
-2026-01-16 07:38:11 INFO: 11.76	7.72	9.78
-2026-01-16 07:38:11 INFO: step 2700: train_loss = 1.947624, dev_score = 0.1176
-2026-01-16 07:38:11 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:38:11 INFO: new model checkpoint saved.
-2026-01-16 07:38:12 INFO: Finished STEP 2720/50000, loss = 1.433867 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:12 INFO: Finished STEP 2740/50000, loss = 0.081805 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:13 INFO: Finished STEP 2760/50000, loss = 2.694524 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:38:14 INFO: Finished STEP 2780/50000, loss = 2.323376 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:14 INFO: Finished STEP 2800/50000, loss = 0.507773 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:38:14 INFO: Evaluating on dev set...
-2026-01-16 07:38:16 INFO: LAS	MLAS	BLEX
-2026-01-16 07:38:16 INFO: 16.05	7.79	9.64
-2026-01-16 07:38:16 INFO: step 2800: train_loss = 2.021911, dev_score = 0.1605
-2026-01-16 07:38:17 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:38:17 INFO: new model checkpoint saved.
-2026-01-16 07:38:17 INFO: Finished STEP 2820/50000, loss = 2.562084 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:18 INFO: Finished STEP 2840/50000, loss = 2.842648 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:38:18 INFO: Finished STEP 2860/50000, loss = 0.741765 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:38:19 INFO: Finished STEP 2880/50000, loss = 1.145487 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:20 INFO: Finished STEP 2900/50000, loss = 1.986015 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:38:20 INFO: Evaluating on dev set...
-2026-01-16 07:38:22 INFO: LAS	MLAS	BLEX
-2026-01-16 07:38:22 INFO: 19.42	11.02	14.05
-2026-01-16 07:38:22 INFO: step 2900: train_loss = 1.980637, dev_score = 0.1942
-2026-01-16 07:38:22 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:38:22 INFO: new best model saved.
-2026-01-16 07:38:23 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:38:23 INFO: new model checkpoint saved.
-2026-01-16 07:38:23 INFO: Finished STEP 2920/50000, loss = 2.065670 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:38:24 INFO: Finished STEP 2940/50000, loss = 1.709635 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:24 INFO: Finished STEP 2960/50000, loss = 3.801945 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:25 INFO: Finished STEP 2980/50000, loss = 1.160876 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:26 INFO: Finished STEP 3000/50000, loss = 3.921757 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:26 INFO: Evaluating on dev set...
-2026-01-16 07:38:28 INFO: LAS	MLAS	BLEX
-2026-01-16 07:38:28 INFO: 12.19	4.47	6.57
-2026-01-16 07:38:28 INFO: step 3000: train_loss = 1.918748, dev_score = 0.1219
-2026-01-16 07:38:29 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:38:29 INFO: new model checkpoint saved.
-2026-01-16 07:38:29 INFO: Finished STEP 3020/50000, loss = 1.314914 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:30 INFO: Finished STEP 3040/50000, loss = 3.862856 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:38:31 INFO: Finished STEP 3060/50000, loss = 2.133690 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:38:31 INFO: Finished STEP 3080/50000, loss = 2.644999 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:38:32 INFO: Finished STEP 3100/50000, loss = 2.620658 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:38:32 INFO: Evaluating on dev set...
-2026-01-16 07:38:34 INFO: LAS	MLAS	BLEX
-2026-01-16 07:38:34 INFO: 9.56	5.70	7.32
-2026-01-16 07:38:34 INFO: step 3100: train_loss = 1.990910, dev_score = 0.0956
-2026-01-16 07:38:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:38:34 INFO: new model checkpoint saved.
-2026-01-16 07:38:35 INFO: Finished STEP 3120/50000, loss = 1.526267 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:38:35 INFO: Finished STEP 3140/50000, loss = 1.378218 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:38:36 INFO: Finished STEP 3160/50000, loss = 1.390367 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:37 INFO: Finished STEP 3180/50000, loss = 1.705234 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:37 INFO: Finished STEP 3200/50000, loss = 1.402759 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:37 INFO: Evaluating on dev set...
-2026-01-16 07:38:40 INFO: LAS	MLAS	BLEX
-2026-01-16 07:38:40 INFO: 17.65	10.33	12.02
-2026-01-16 07:38:40 INFO: step 3200: train_loss = 2.082199, dev_score = 0.1765
-2026-01-16 07:38:40 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:38:40 INFO: new model checkpoint saved.
-2026-01-16 07:38:41 INFO: Finished STEP 3220/50000, loss = 1.167270 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:42 INFO: Finished STEP 3240/50000, loss = 3.461316 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:38:42 INFO: Finished STEP 3260/50000, loss = 2.926430 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:43 INFO: Finished STEP 3280/50000, loss = 3.029675 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:44 INFO: Finished STEP 3300/50000, loss = 2.962304 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:44 INFO: Evaluating on dev set...
-2026-01-16 07:38:46 INFO: LAS	MLAS	BLEX
-2026-01-16 07:38:46 INFO: 21.14	14.06	15.91
-2026-01-16 07:38:46 INFO: step 3300: train_loss = 2.159422, dev_score = 0.2114
-2026-01-16 07:38:46 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:38:46 INFO: new best model saved.
-2026-01-16 07:38:47 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:38:47 INFO: new model checkpoint saved.
-2026-01-16 07:38:47 INFO: Finished STEP 3320/50000, loss = 1.153689 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:38:48 INFO: Finished STEP 3340/50000, loss = 6.680528 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:49 INFO: Finished STEP 3360/50000, loss = 5.256881 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:38:49 INFO: Finished STEP 3380/50000, loss = 2.249988 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:50 INFO: Finished STEP 3400/50000, loss = 2.187397 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:38:50 INFO: Evaluating on dev set...
-2026-01-16 07:38:52 INFO: LAS	MLAS	BLEX
-2026-01-16 07:38:52 INFO: 10.85	9.14	11.20
-2026-01-16 07:38:52 INFO: step 3400: train_loss = 1.977379, dev_score = 0.1085
-2026-01-16 07:38:52 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:38:52 INFO: new model checkpoint saved.
-2026-01-16 07:38:53 INFO: Finished STEP 3420/50000, loss = 1.144123 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:38:53 INFO: Finished STEP 3440/50000, loss = 1.048997 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:54 INFO: Finished STEP 3460/50000, loss = 2.239283 (0.033 sec/batch), lr: 0.003000
-2026-01-16 07:38:55 INFO: Finished STEP 3480/50000, loss = 1.407322 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:38:55 INFO: Finished STEP 3500/50000, loss = 1.793783 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:38:55 INFO: Evaluating on dev set...
-2026-01-16 07:38:57 INFO: LAS	MLAS	BLEX
-2026-01-16 07:38:57 INFO: 25.55	14.33	15.18
-2026-01-16 07:38:57 INFO: step 3500: train_loss = 1.776203, dev_score = 0.2555
-2026-01-16 07:38:58 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:38:58 INFO: new best model saved.
-2026-01-16 07:38:58 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:38:58 INFO: new model checkpoint saved.
-2026-01-16 07:38:59 INFO: Finished STEP 3520/50000, loss = 1.173083 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:38:59 INFO: Finished STEP 3540/50000, loss = 2.557373 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:39:00 INFO: Finished STEP 3560/50000, loss = 0.362348 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:39:01 INFO: Finished STEP 3580/50000, loss = 0.959906 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:01 INFO: Finished STEP 3600/50000, loss = 0.376063 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:39:01 INFO: Evaluating on dev set...
-2026-01-16 07:39:03 INFO: LAS	MLAS	BLEX
-2026-01-16 07:39:03 INFO: 10.60	5.75	9.91
-2026-01-16 07:39:03 INFO: step 3600: train_loss = 1.918432, dev_score = 0.1060
-2026-01-16 07:39:04 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:39:04 INFO: new model checkpoint saved.
-2026-01-16 07:39:04 INFO: Finished STEP 3620/50000, loss = 0.217418 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:05 INFO: Finished STEP 3640/50000, loss = 1.392662 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:05 INFO: Finished STEP 3660/50000, loss = 2.835163 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:06 INFO: Finished STEP 3680/50000, loss = 2.799474 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:07 INFO: Finished STEP 3700/50000, loss = 0.710460 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:07 INFO: Evaluating on dev set...
-2026-01-16 07:39:08 INFO: LAS	MLAS	BLEX
-2026-01-16 07:39:08 INFO: 13.11	10.84	12.64
-2026-01-16 07:39:08 INFO: step 3700: train_loss = 1.913150, dev_score = 0.1311
-2026-01-16 07:39:09 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:39:09 INFO: new model checkpoint saved.
-2026-01-16 07:39:10 INFO: Finished STEP 3720/50000, loss = 2.065038 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:10 INFO: Finished STEP 3740/50000, loss = 1.423430 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:39:11 INFO: Finished STEP 3760/50000, loss = 0.174135 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:39:11 INFO: Finished STEP 3780/50000, loss = 0.994005 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:12 INFO: Finished STEP 3800/50000, loss = 2.113207 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:12 INFO: Evaluating on dev set...
-2026-01-16 07:39:14 INFO: LAS	MLAS	BLEX
-2026-01-16 07:39:14 INFO: 11.70	9.55	12.36
-2026-01-16 07:39:14 INFO: step 3800: train_loss = 1.940419, dev_score = 0.1170
-2026-01-16 07:39:14 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:39:14 INFO: new model checkpoint saved.
-2026-01-16 07:39:15 INFO: Finished STEP 3820/50000, loss = 1.886823 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:16 INFO: Finished STEP 3840/50000, loss = 0.537330 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:39:16 INFO: Finished STEP 3860/50000, loss = 1.127199 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:17 INFO: Finished STEP 3880/50000, loss = 1.348202 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:17 INFO: Finished STEP 3900/50000, loss = 2.103697 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:39:17 INFO: Evaluating on dev set...
-2026-01-16 07:39:19 INFO: LAS	MLAS	BLEX
-2026-01-16 07:39:19 INFO: 19.36	13.16	14.37
-2026-01-16 07:39:19 INFO: step 3900: train_loss = 2.127213, dev_score = 0.1936
-2026-01-16 07:39:20 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:39:20 INFO: new model checkpoint saved.
-2026-01-16 07:39:20 INFO: Finished STEP 3920/50000, loss = 2.552736 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:39:21 INFO: Finished STEP 3940/50000, loss = 0.713954 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:22 INFO: Finished STEP 3960/50000, loss = 0.829729 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:22 INFO: Finished STEP 3980/50000, loss = 2.967856 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:23 INFO: Finished STEP 4000/50000, loss = 2.044243 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:39:23 INFO: Evaluating on dev set...
-2026-01-16 07:39:25 INFO: LAS	MLAS	BLEX
-2026-01-16 07:39:25 INFO: 13.91	7.27	8.25
-2026-01-16 07:39:25 INFO: step 4000: train_loss = 1.918335, dev_score = 0.1391
-2026-01-16 07:39:25 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:39:25 INFO: new model checkpoint saved.
-2026-01-16 07:39:26 INFO: Finished STEP 4020/50000, loss = 0.570849 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:26 INFO: Finished STEP 4040/50000, loss = 0.004094 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:27 INFO: Finished STEP 4060/50000, loss = 0.615223 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:39:28 INFO: Finished STEP 4080/50000, loss = 1.598375 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:39:28 INFO: Finished STEP 4100/50000, loss = 1.763149 (0.033 sec/batch), lr: 0.003000
-2026-01-16 07:39:28 INFO: Evaluating on dev set...
-2026-01-16 07:39:30 INFO: LAS	MLAS	BLEX
-2026-01-16 07:39:30 INFO: 17.22	6.34	9.61
-2026-01-16 07:39:30 INFO: step 4100: train_loss = 1.769413, dev_score = 0.1722
-2026-01-16 07:39:31 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:39:31 INFO: new model checkpoint saved.
-2026-01-16 07:39:31 INFO: Finished STEP 4120/50000, loss = 4.137807 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:39:32 INFO: Finished STEP 4140/50000, loss = 3.163526 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:32 INFO: Finished STEP 4160/50000, loss = 7.217985 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:33 INFO: Finished STEP 4180/50000, loss = 1.032177 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:39:34 INFO: Finished STEP 4200/50000, loss = 0.475087 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:34 INFO: Evaluating on dev set...
-2026-01-16 07:39:36 INFO: LAS	MLAS	BLEX
-2026-01-16 07:39:36 INFO: 10.11	6.28	11.64
-2026-01-16 07:39:36 INFO: step 4200: train_loss = 1.876384, dev_score = 0.1011
-2026-01-16 07:39:36 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:39:36 INFO: new model checkpoint saved.
-2026-01-16 07:39:37 INFO: Finished STEP 4220/50000, loss = 1.246820 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:37 INFO: Finished STEP 4240/50000, loss = 1.793286 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:38 INFO: Finished STEP 4260/50000, loss = 2.043184 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:39:39 INFO: Finished STEP 4280/50000, loss = 2.374679 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:39 INFO: Finished STEP 4300/50000, loss = 1.198715 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:39:39 INFO: Evaluating on dev set...
-2026-01-16 07:39:41 INFO: LAS	MLAS	BLEX
-2026-01-16 07:39:41 INFO: 12.13	4.51	8.45
-2026-01-16 07:39:41 INFO: step 4300: train_loss = 1.807878, dev_score = 0.1213
-2026-01-16 07:39:42 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:39:42 INFO: new model checkpoint saved.
-2026-01-16 07:39:42 INFO: Finished STEP 4320/50000, loss = 1.510360 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:43 INFO: Finished STEP 4340/50000, loss = 2.409623 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:43 INFO: Finished STEP 4360/50000, loss = 1.254725 (0.031 sec/batch), lr: 0.003000
-2026-01-16 07:39:44 INFO: Finished STEP 4380/50000, loss = 7.248045 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:45 INFO: Finished STEP 4400/50000, loss = 2.985596 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:45 INFO: Evaluating on dev set...
-2026-01-16 07:39:46 INFO: LAS	MLAS	BLEX
-2026-01-16 07:39:46 INFO: 11.95	7.40	9.70
-2026-01-16 07:39:46 INFO: step 4400: train_loss = 1.787297, dev_score = 0.1195
-2026-01-16 07:39:47 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2026-01-16 07:39:47 INFO: new model checkpoint saved.
-2026-01-16 07:39:47 INFO: Finished STEP 4420/50000, loss = 1.253725 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:48 INFO: Finished STEP 4440/50000, loss = 2.283563 (0.029 sec/batch), lr: 0.003000
-2026-01-16 07:39:49 INFO: Finished STEP 4460/50000, loss = 2.580783 (0.030 sec/batch), lr: 0.003000
-2026-01-16 07:39:49 INFO: Finished STEP 4480/50000, loss = 1.721368 (0.032 sec/batch), lr: 0.003000
-2026-01-16 07:39:50 INFO: Finished STEP 4500/50000, loss = 1.226920 (0.034 sec/batch), lr: 0.003000
-2026-01-16 07:39:50 INFO: Evaluating on dev set...
-2026-01-16 07:39:52 INFO: LAS	MLAS	BLEX
-2026-01-16 07:39:52 INFO: 18.81	11.17	13.67
-2026-01-16 07:39:52 INFO: step 4500: train_loss = 2.092089, dev_score = 0.1881
-2026-01-16 07:39:52 INFO: Training ended with 4500 steps.
-2026-01-16 07:39:52 INFO: Best dev F1 = 25.55, at iteration = 3500
-2026-01-16 07:39:53 INFO: Running dev depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '8', '--dropout', '0.15']
-2026-01-16 07:39:53 INFO: Running parser in predict mode
-2026-01-16 07:39:53 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:39:54 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2026-01-16 07:39:54 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:39:54 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2026-01-16 07:39:54 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:39:55 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2026-01-16 07:39:55 INFO: Loading data with batch size 8...
-2026-01-16 07:39:55 DEBUG: 34 batches created.
-2026-01-16 07:39:57 INFO: F1 scores for each dependency:
-  Note that unlabeled attachment errors hurt the labeled attachment scores
-         acl: p 0.0000 r 0.0000 f1 0.0000 (14 actual)
-   acl:relcl: p 0.0000 r 0.0000 f1 0.0000 (32 actual)
-       advcl: p 0.0000 r 0.0000 f1 0.0000 (26 actual)
- advcl:relcl: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-      advmod: p 0.1391 r 0.5081 f1 0.2184 (124 actual)
-        amod: p 0.1538 r 0.0194 f1 0.0345 (103 actual)
-       appos: p 0.0000 r 0.0000 f1 0.0000 (12 actual)
-         aux: p 0.0000 r 0.0000 f1 0.0000 (42 actual)
-    aux:pass: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-        case: p 0.6946 r 0.6784 f1 0.6864 (171 actual)
-          cc: p 0.0000 r 0.0000 f1 0.0000 (67 actual)
-       ccomp: p 0.0000 r 0.0000 f1 0.0000 (17 actual)
-compound:prt: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
-        conj: p 0.0000 r 0.0000 f1 0.0000 (64 actual)
-         cop: p 0.0000 r 0.0000 f1 0.0000 (16 actual)
-       csubj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-         dep: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-         det: p 0.7222 r 0.7800 f1 0.7500 (100 actual)
-   discourse: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
-        expl: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-       fixed: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
-        flat: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-   flat:name: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
-    goeswith: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-        iobj: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
-        mark: p 0.0000 r 0.0000 f1 0.0000 (68 actual)
-        nmod: p 0.0000 r 0.0000 f1 0.0000 (55 actual)
-   nmod:poss: p 0.3750 r 0.0484 f1 0.0857 (62 actual)
-       nsubj: p 0.1684 r 0.5294 f1 0.2556 (119 actual)
-  nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (10 actual)
-      nummod: p 0.0000 r 0.0000 f1 0.0000 (10 actual)
-         obj: p 0.0517 r 0.1481 f1 0.0767 (81 actual)
-         obl: p 0.0270 r 0.0080 f1 0.0123 (125 actual)
-   obl:agent: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-   parataxis: p 0.0000 r 0.0000 f1 0.0000 (13 actual)
-       punct: p 0.3245 r 0.3333 f1 0.3288 (183 actual)
-  reparandum: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-        root: p 0.4412 r 0.4412 f1 0.4412 (34 actual)
-    vocative: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
-       xcomp: p 0.0000 r 0.0000 f1 0.0000 (28 actual)
-2026-01-16 07:39:57 INFO: LAS	MLAS	BLEX
-2026-01-16 07:39:57 INFO: 25.55	14.33	15.18
-2026-01-16 07:39:57 INFO: Parser score:
-2026-01-16 07:39:57 INFO: sv_diachronic 25.55
-2026-01-16 07:39:57 INFO: Finished running dev set on
-UD_Swedish-diachronic
-  UAS   LAS  CLAS  MLAS  BLEX
-49.08 25.55 15.18 14.33 15.18
-2026-01-16 07:39:57 INFO: Running test depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '8', '--dropout', '0.15']
-2026-01-16 07:39:57 INFO: Running parser in predict mode
-2026-01-16 07:39:57 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
-2026-01-16 07:39:59 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2026-01-16 07:39:59 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:39:59 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2026-01-16 07:39:59 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2026-01-16 07:39:59 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2026-01-16 07:39:59 INFO: Loading data with batch size 8...
-2026-01-16 07:39:59 DEBUG: 74 batches created.
-2026-01-16 07:40:02 INFO: F1 scores for each dependency:
-  Note that unlabeled attachment errors hurt the labeled attachment scores
-         acl: p 0.0000 r 0.0000 f1 0.0000 (21 actual)
-   acl:cleft: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-   acl:relcl: p 0.0000 r 0.0000 f1 0.0000 (50 actual)
-       advcl: p 0.0000 r 0.0000 f1 0.0000 (39 actual)
-      advmod: p 0.1618 r 0.5858 f1 0.2535 (169 actual)
-        amod: p 0.1698 r 0.0570 f1 0.0853 (158 actual)
-       appos: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
-         aux: p 0.0000 r 0.0000 f1 0.0000 (53 actual)
-    aux:pass: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-        case: p 0.7148 r 0.7868 f1 0.7491 (258 actual)
-          cc: p 0.0000 r 0.0000 f1 0.0000 (101 actual)
-       ccomp: p 0.0000 r 0.0000 f1 0.0000 (20 actual)
-compound:prt: p 0.0000 r 0.0000 f1 0.0000 (14 actual)
-        conj: p 0.0000 r 0.0000 f1 0.0000 (106 actual)
-         cop: p 0.0000 r 0.0000 f1 0.0000 (33 actual)
-       csubj: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-         det: p 0.6382 r 0.7462 f1 0.6879 (130 actual)
-   discourse: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-  dislocated: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-        expl: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
-     expl:pv: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-       fixed: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-   flat:name: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
-        iobj: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
-        mark: p 0.0000 r 0.0000 f1 0.0000 (97 actual)
-        nmod: p 0.0000 r 0.0000 f1 0.0000 (62 actual)
-   nmod:poss: p 0.0588 r 0.0101 f1 0.0172 (99 actual)
-       nsubj: p 0.1690 r 0.4719 f1 0.2489 (178 actual)
-  nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (20 actual)
-         obj: p 0.0508 r 0.1532 f1 0.0763 (124 actual)
-         obl: p 0.0000 r 0.0000 f1 0.0000 (194 actual)
-   obl:agent: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
-      orphan: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-   parataxis: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
-       punct: p 0.3435 r 0.3435 f1 0.3435 (294 actual)
-        root: p 0.4189 r 0.4189 f1 0.4189 (74 actual)
-    vocative: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-       xcomp: p 0.0000 r 0.0000 f1 0.0000 (55 actual)
-2026-01-16 07:40:02 INFO: LAS	MLAS	BLEX
-2026-01-16 07:40:02 INFO: 27.02	15.21	16.05
-2026-01-16 07:40:02 INFO: Parser score:
-2026-01-16 07:40:02 INFO: sv_diachronic 27.02
-2026-01-16 07:40:03 INFO: Finished running test set on
-UD_Swedish-diachronic
-  UAS   LAS  CLAS  MLAS  BLEX
-49.63 27.02 16.05 15.21 16.05
-DONE.
-Full log saved to: logs/log_conll17.pt_sv_diachron_20260116_073035.txt
-Symlink updated: logs/latest.txt → log_conll17.pt_sv_diachron_20260116_073035.txt

logs/log_conll17.pt_sv_diachron_de_lit_20251203_154509.txt

@@ -1,740 +0,0 @@
-=== LOGFILE: logs/log_conll17.pt_sv_diachron_de_lit_20251203_154509.txt ===
-Language codes: sv diachron de_lit
-Using pretrained model: conll17.pt
-
-Running: python prepare-train-val-test.py sv diachron de_lit
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/ucxn_ud_swedish-talbanken.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
-Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-de_lit/de_lit-ud-test.conllu
-Including DigPhil MACHINE in TRAIN (minus gold)…
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
-Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
-Cleaning TRAIN...
-[REMOVED] sent_id=6  ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
-[REMOVED] sent_id=7_8  ERRORS=['Multiple roots found: [5, 10]']
-[REMOVED] sent_id=30_31  ERRORS=['Multiple roots found: [3, 18]']
-[REMOVED] sent_id=35  ERRORS=['Line 36: Invalid token ID or head']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=2_3  ERRORS=['Multiple roots found: [7, 20]']
-[REMOVED] sent_id=8_9  ERRORS=['Multiple roots found: [24, 57]']
-[REMOVED] sent_id=12_13  ERRORS=['Multiple roots found: [11, 16]']
-[REMOVED] sent_id=124_split2  ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
-[REMOVED] sent_id=396  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=416  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=589  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=909  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=912  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=3_split1  ERRORS=['Multiple roots found: [4, 15, 17]']
-[REMOVED] sent_id=3_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
-[REMOVED] sent_id=3_4  ERRORS=['Multiple roots found: [1, 5]']
-[REMOVED] sent_id=5_6  ERRORS=['Multiple roots found: [3, 24]']
-[REMOVED] sent_id=11_12_13  ERRORS=['Multiple roots found: [5, 17, 25]']
-[REMOVED] sent_id=119  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=188  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=223  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=268  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=325  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=388  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=399  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=475  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=505  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=520  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=562  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=711  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=731  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=867  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=884  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=923  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=939  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1086  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1179  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1251  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1345  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1459  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1656  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=1669  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=87_88  ERRORS=['Multiple roots found: [3, 6]']
-[REMOVED] sent_id=65_split2_66_split2  ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
-[REMOVED] sent_id=25  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=136  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=208  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=230  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=245  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=276  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=320  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=366  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=519  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=569  ERRORS=['Token 2: Missing form']
-[REMOVED] sent_id=50_split2  ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
-[REMOVED] sent_id=53_54  ERRORS=['Multiple roots found: [27, 91]']
-[REMOVED] sent_id=55_56_57  ERRORS=['Multiple roots found: [2, 4, 13]']
-[REMOVED] sent_id=17_split1  ERRORS=['Multiple roots found: [2, 14, 17]']
-[REMOVED] sent_id=17_split2  ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
-[REMOVED] sent_id=19_split1  ERRORS=['Multiple roots found: [3, 31]']
-Cleaning DEV...
-[REMOVED] sent_id=33  ERRORS=['Token 15: Missing deprel']
-Cleaning TEST...
-Writing TRAIN → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (68172 valid sentences)
-Writing DEV → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (9 valid sentences)
-Writing TEST → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (99 valid sentences)
-Done.
-Sourcing scripts/config_alvis.sh
-Running stanza dataset preparation…
-2025-12-03 15:45:18 INFO: Datasets program called with:
-/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 15:45:18 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
-
-2025-12-03 15:45:18 INFO: Downloaded file to /cephyr/users/cleland/Alvis/stanza_resources/resources.json
-2025-12-03 15:45:18 DEBUG: Processing parameter "processors"...
-2025-12-03 15:45:18 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
-2025-12-03 15:45:18 INFO: Downloading these customized packages for language: sv (Swedish)...
-=======================
-| Processor | Package |
------------------------
-=======================
-
-2025-12-03 15:45:18 INFO: Finished downloading models and saved to /cephyr/users/cleland/Alvis/stanza_resources
-2025-12-03 15:45:18 INFO: Using tagger model in /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt for sv_diachronic
-2025-12-03 15:45:18 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2025-12-03 15:45:18 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-Augmented 188 quotes: Counter({'«»': 23, '»«': 22, '″″': 22, '""': 20, '““': 20, '„”': 19, '「」': 18, '《》': 16, '„“': 15, '””': 13})
-2025-12-03 15:45:21 INFO: Running tagger to retag /local/tmp.5440223/tmplttwhhzq/sv_diachronic.train.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmplttwhhzq/sv_diachronic.train.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu']
-2025-12-03 15:45:22 INFO: Running tagger in predict mode
-2025-12-03 15:45:22 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2025-12-03 15:45:23 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 15:45:23 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 15:45:23 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 15:45:23 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 15:45:24 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 15:45:26 INFO: Loading data with batch size 250...
-2025-12-03 15:46:06 INFO: Start evaluation...
-2025-12-03 15:49:44 INFO: UPOS	XPOS	UFeats	AllTags
-2025-12-03 15:49:44 INFO: 96.83	85.29	95.46	84.47
-2025-12-03 15:49:44 INFO: POS Tagger score: sv_diachronic 84.47
-2025-12-03 15:49:45 INFO: Running tagger to retag /local/tmp.5440223/tmplttwhhzq/sv_diachronic.dev.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmplttwhhzq/sv_diachronic.dev.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu']
-2025-12-03 15:49:45 INFO: Running tagger in predict mode
-2025-12-03 15:49:45 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2025-12-03 15:49:47 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 15:49:47 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 15:49:47 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 15:49:47 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 15:49:47 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 15:49:47 INFO: Loading data with batch size 250...
-2025-12-03 15:49:47 INFO: Start evaluation...
-2025-12-03 15:49:47 INFO: UPOS	XPOS	UFeats	AllTags
-2025-12-03 15:49:47 INFO: 93.32	90.84	93.32	85.64
-2025-12-03 15:49:47 INFO: POS Tagger score: sv_diachronic 85.64
-2025-12-03 15:49:47 INFO: Running tagger to retag /local/tmp.5440223/tmplttwhhzq/sv_diachronic.test.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
-  Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmplttwhhzq/sv_diachronic.test.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu']
-2025-12-03 15:49:47 INFO: Running tagger in predict mode
-2025-12-03 15:49:47 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
-2025-12-03 15:49:49 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 15:49:49 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 15:49:49 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 15:49:49 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 15:49:49 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 15:49:49 INFO: Loading data with batch size 250...
-2025-12-03 15:49:49 INFO: Start evaluation...
-2025-12-03 15:49:50 INFO: UPOS	XPOS	UFeats	AllTags
-2025-12-03 15:49:50 INFO: 93.14	96.78	95.32	90.28
-2025-12-03 15:49:50 INFO: POS Tagger score: sv_diachronic 90.28
-Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /local/tmp.5440223/tmplttwhhzq/sv_diachronic.train.gold.conllu
-Swapped 'w1, w2' for 'w1 ,w2' 141 times
-Added 573 new sentences with asdf, zzzz -> asdf,zzzz
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /local/tmp.5440223/tmplttwhhzq/sv_diachronic.dev.gold.conllu
-Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /local/tmp.5440223/tmplttwhhzq/sv_diachronic.test.gold.conllu
-Running stanza dependency parser training…
-2025-12-03 15:50:04 INFO: Training program called with:
-/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt --batch_size 32 --dropout 0.33
-2025-12-03 15:50:04 DEBUG: UD_Swedish-diachronic: sv_diachronic
-2025-12-03 15:50:04 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2025-12-03 15:50:04 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2025-12-03 15:50:04 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_charlm_parser.pt does not exist, training new model
-2025-12-03 15:50:04 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
-2025-12-03 15:50:04 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
-2025-12-03 15:50:04 INFO: Running train depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--train_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--batch_size', '5000', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'train', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
-2025-12-03 15:50:04 INFO: Running parser in train mode
-2025-12-03 15:50:04 INFO: Using pretrained contextualized char embedding
-2025-12-03 15:50:04 INFO: Loading data with batch size 32...
-2025-12-03 15:50:11 INFO: Train File /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu, Data Size: 68745
-2025-12-03 15:50:11 INFO: Original data size: 68745
-2025-12-03 15:50:12 INFO: Augmented data size: 69046
-2025-12-03 15:50:31 WARNING: sv_diachronic is not a known dataset.  Examining the data to choose which xpos vocab to use
-2025-12-03 15:50:31 INFO: Original length = 69046
-2025-12-03 15:50:31 INFO: Filtered length = 69046
-2025-12-03 15:50:47 WARNING: Chose XPOSDescription(xpos_type=<XPOSType.XPOS: 1>, sep='|') for the xpos factory for sv_diachronic
-2025-12-03 15:50:55 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 15:51:09 DEBUG: 40457 batches created.
-2025-12-03 15:51:09 DEBUG: 9 batches created.
-2025-12-03 15:51:09 INFO: Training parser...
-2025-12-03 15:51:09 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 15:51:09 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 15:51:09 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 15:51:10 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 15:51:17 INFO: Finished STEP 20/50000, loss = 6.791992 (0.200 sec/batch), lr: 0.003000
-2025-12-03 15:51:21 INFO: Finished STEP 40/50000, loss = 5.211706 (0.167 sec/batch), lr: 0.003000
-2025-12-03 15:51:24 INFO: Finished STEP 60/50000, loss = 5.668643 (0.147 sec/batch), lr: 0.003000
-2025-12-03 15:51:27 INFO: Finished STEP 80/50000, loss = 4.197357 (0.151 sec/batch), lr: 0.003000
-2025-12-03 15:51:30 INFO: Finished STEP 100/50000, loss = 5.356755 (0.129 sec/batch), lr: 0.003000
-2025-12-03 15:51:30 INFO: Evaluating on dev set...
-2025-12-03 15:51:31 INFO: LAS	MLAS	BLEX
-2025-12-03 15:51:31 INFO: 39.85	26.44	33.26
-2025-12-03 15:51:31 INFO: step 100: train_loss = 8.587716, dev_score = 0.3985
-2025-12-03 15:51:31 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 15:51:31 INFO: new best model saved.
-2025-12-03 15:51:32 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:51:32 INFO: new model checkpoint saved.
-2025-12-03 15:51:34 INFO: Finished STEP 120/50000, loss = 4.996027 (0.120 sec/batch), lr: 0.003000
-2025-12-03 15:51:37 INFO: Finished STEP 140/50000, loss = 5.125143 (0.131 sec/batch), lr: 0.003000
-2025-12-03 15:51:40 INFO: Finished STEP 160/50000, loss = 4.139769 (0.138 sec/batch), lr: 0.003000
-2025-12-03 15:51:42 INFO: Finished STEP 180/50000, loss = 4.071211 (0.131 sec/batch), lr: 0.003000
-2025-12-03 15:51:45 INFO: Finished STEP 200/50000, loss = 4.302421 (0.121 sec/batch), lr: 0.003000
-2025-12-03 15:51:45 INFO: Evaluating on dev set...
-2025-12-03 15:51:45 INFO: LAS	MLAS	BLEX
-2025-12-03 15:51:45 INFO: 50.00	35.44	40.93
-2025-12-03 15:51:45 INFO: step 200: train_loss = 4.776229, dev_score = 0.5000
-2025-12-03 15:51:45 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 15:51:45 INFO: new best model saved.
-2025-12-03 15:51:46 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:51:46 INFO: new model checkpoint saved.
-2025-12-03 15:51:48 INFO: Finished STEP 220/50000, loss = 4.349181 (0.124 sec/batch), lr: 0.003000
-2025-12-03 15:51:51 INFO: Finished STEP 240/50000, loss = 5.862440 (0.125 sec/batch), lr: 0.003000
-2025-12-03 15:51:53 INFO: Finished STEP 260/50000, loss = 4.765276 (0.118 sec/batch), lr: 0.003000
-2025-12-03 15:51:56 INFO: Finished STEP 280/50000, loss = 4.617913 (0.116 sec/batch), lr: 0.003000
-2025-12-03 15:51:58 INFO: Finished STEP 300/50000, loss = 5.353960 (0.113 sec/batch), lr: 0.003000
-2025-12-03 15:51:58 INFO: Evaluating on dev set...
-2025-12-03 15:51:58 INFO: LAS	MLAS	BLEX
-2025-12-03 15:51:58 INFO: 51.49	38.56	43.22
-2025-12-03 15:51:58 INFO: step 300: train_loss = 4.421819, dev_score = 0.5149
-2025-12-03 15:51:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 15:51:59 INFO: new best model saved.
-2025-12-03 15:51:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:51:59 INFO: new model checkpoint saved.
-2025-12-03 15:52:02 INFO: Finished STEP 320/50000, loss = 5.069269 (0.121 sec/batch), lr: 0.003000
-2025-12-03 15:52:04 INFO: Finished STEP 340/50000, loss = 3.614627 (0.112 sec/batch), lr: 0.003000
-2025-12-03 15:52:06 INFO: Finished STEP 360/50000, loss = 3.878788 (0.115 sec/batch), lr: 0.003000
-2025-12-03 15:52:08 INFO: Finished STEP 380/50000, loss = 4.367438 (0.100 sec/batch), lr: 0.003000
-2025-12-03 15:52:10 INFO: Finished STEP 400/50000, loss = 4.081972 (0.107 sec/batch), lr: 0.003000
-2025-12-03 15:52:10 INFO: Evaluating on dev set...
-2025-12-03 15:52:11 INFO: LAS	MLAS	BLEX
-2025-12-03 15:52:11 INFO: 55.69	43.55	46.51
-2025-12-03 15:52:11 INFO: step 400: train_loss = 4.280971, dev_score = 0.5569
-2025-12-03 15:52:11 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 15:52:11 INFO: new best model saved.
-2025-12-03 15:52:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:52:12 INFO: new model checkpoint saved.
-2025-12-03 15:52:14 INFO: Finished STEP 420/50000, loss = 3.892664 (0.110 sec/batch), lr: 0.003000
-2025-12-03 15:52:16 INFO: Finished STEP 440/50000, loss = 3.998952 (0.100 sec/batch), lr: 0.003000
-2025-12-03 15:52:18 INFO: Finished STEP 460/50000, loss = 5.833274 (0.112 sec/batch), lr: 0.003000
-2025-12-03 15:52:20 INFO: Finished STEP 480/50000, loss = 3.561453 (0.106 sec/batch), lr: 0.003000
-2025-12-03 15:52:22 INFO: Finished STEP 500/50000, loss = 3.267099 (0.106 sec/batch), lr: 0.003000
-2025-12-03 15:52:22 INFO: Evaluating on dev set...
-2025-12-03 15:52:23 INFO: LAS	MLAS	BLEX
-2025-12-03 15:52:23 INFO: 55.45	44.30	48.95
-2025-12-03 15:52:23 INFO: step 500: train_loss = 4.074671, dev_score = 0.5545
-2025-12-03 15:52:24 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:52:24 INFO: new model checkpoint saved.
-2025-12-03 15:52:26 INFO: Finished STEP 520/50000, loss = 3.987905 (0.101 sec/batch), lr: 0.003000
-2025-12-03 15:52:28 INFO: Finished STEP 540/50000, loss = 4.040174 (0.105 sec/batch), lr: 0.003000
-2025-12-03 15:52:30 INFO: Finished STEP 560/50000, loss = 5.425183 (0.096 sec/batch), lr: 0.003000
-2025-12-03 15:52:32 INFO: Finished STEP 580/50000, loss = 3.171549 (0.098 sec/batch), lr: 0.003000
-2025-12-03 15:52:34 INFO: Finished STEP 600/50000, loss = 5.054599 (0.087 sec/batch), lr: 0.003000
-2025-12-03 15:52:34 INFO: Evaluating on dev set...
-2025-12-03 15:52:34 INFO: LAS	MLAS	BLEX
-2025-12-03 15:52:34 INFO: 59.41	50.85	53.39
-2025-12-03 15:52:34 INFO: step 600: train_loss = 3.989031, dev_score = 0.5941
-2025-12-03 15:52:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 15:52:34 INFO: new best model saved.
-2025-12-03 15:52:35 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:52:35 INFO: new model checkpoint saved.
-2025-12-03 15:52:37 INFO: Finished STEP 620/50000, loss = 5.185298 (0.107 sec/batch), lr: 0.003000
-2025-12-03 15:52:39 INFO: Finished STEP 640/50000, loss = 3.282787 (0.098 sec/batch), lr: 0.003000
-2025-12-03 15:52:41 INFO: Finished STEP 660/50000, loss = 7.106239 (0.095 sec/batch), lr: 0.003000
-2025-12-03 15:52:43 INFO: Finished STEP 680/50000, loss = 3.663255 (0.099 sec/batch), lr: 0.003000
-2025-12-03 15:52:45 INFO: Finished STEP 700/50000, loss = 4.011485 (0.098 sec/batch), lr: 0.003000
-2025-12-03 15:52:45 INFO: Evaluating on dev set...
-2025-12-03 15:52:46 INFO: LAS	MLAS	BLEX
-2025-12-03 15:52:46 INFO: 59.16	46.74	52.21
-2025-12-03 15:52:46 INFO: step 700: train_loss = 3.973323, dev_score = 0.5916
-2025-12-03 15:52:46 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:52:46 INFO: new model checkpoint saved.
-2025-12-03 15:52:48 INFO: Finished STEP 720/50000, loss = 4.034673 (0.094 sec/batch), lr: 0.003000
-2025-12-03 15:52:50 INFO: Finished STEP 740/50000, loss = 4.719307 (0.099 sec/batch), lr: 0.003000
-2025-12-03 15:52:52 INFO: Finished STEP 760/50000, loss = 3.175733 (0.089 sec/batch), lr: 0.003000
-2025-12-03 15:52:54 INFO: Finished STEP 780/50000, loss = 3.158538 (0.098 sec/batch), lr: 0.003000
-2025-12-03 15:52:56 INFO: Finished STEP 800/50000, loss = 2.881503 (0.099 sec/batch), lr: 0.003000
-2025-12-03 15:52:56 INFO: Evaluating on dev set...
-2025-12-03 15:52:56 INFO: LAS	MLAS	BLEX
-2025-12-03 15:52:56 INFO: 57.92	46.19	49.15
-2025-12-03 15:52:56 INFO: step 800: train_loss = 3.763542, dev_score = 0.5792
-2025-12-03 15:52:57 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:52:57 INFO: new model checkpoint saved.
-2025-12-03 15:52:59 INFO: Finished STEP 820/50000, loss = 4.045841 (0.087 sec/batch), lr: 0.003000
-2025-12-03 15:53:01 INFO: Finished STEP 840/50000, loss = 3.735970 (0.088 sec/batch), lr: 0.003000
-2025-12-03 15:53:02 INFO: Finished STEP 860/50000, loss = 4.020472 (0.080 sec/batch), lr: 0.003000
-2025-12-03 15:53:04 INFO: Finished STEP 880/50000, loss = 5.290604 (0.091 sec/batch), lr: 0.003000
-2025-12-03 15:53:06 INFO: Finished STEP 900/50000, loss = 3.524802 (0.089 sec/batch), lr: 0.003000
-2025-12-03 15:53:06 INFO: Evaluating on dev set...
-2025-12-03 15:53:07 INFO: LAS	MLAS	BLEX
-2025-12-03 15:53:07 INFO: 56.93	45.67	49.47
-2025-12-03 15:53:07 INFO: step 900: train_loss = 3.948068, dev_score = 0.5693
-2025-12-03 15:53:07 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:53:07 INFO: new model checkpoint saved.
-2025-12-03 15:53:09 INFO: Finished STEP 920/50000, loss = 3.348033 (0.088 sec/batch), lr: 0.003000
-2025-12-03 15:53:11 INFO: Finished STEP 940/50000, loss = 3.860394 (0.091 sec/batch), lr: 0.003000
-2025-12-03 15:53:13 INFO: Finished STEP 960/50000, loss = 3.459932 (0.082 sec/batch), lr: 0.003000
-2025-12-03 15:53:15 INFO: Finished STEP 980/50000, loss = 4.335596 (0.085 sec/batch), lr: 0.003000
-2025-12-03 15:53:17 INFO: Finished STEP 1000/50000, loss = 3.108686 (0.089 sec/batch), lr: 0.003000
-2025-12-03 15:53:17 INFO: Evaluating on dev set...
-2025-12-03 15:53:17 INFO: LAS	MLAS	BLEX
-2025-12-03 15:53:17 INFO: 60.40	51.27	53.81
-2025-12-03 15:53:17 INFO: step 1000: train_loss = 3.976391, dev_score = 0.6040
-2025-12-03 15:53:17 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 15:53:17 INFO: new best model saved.
-2025-12-03 15:53:18 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:53:18 INFO: new model checkpoint saved.
-2025-12-03 15:53:20 INFO: Finished STEP 1020/50000, loss = 6.081649 (0.089 sec/batch), lr: 0.003000
-2025-12-03 15:53:21 INFO: Finished STEP 1040/50000, loss = 3.454382 (0.089 sec/batch), lr: 0.003000
-2025-12-03 15:53:23 INFO: Finished STEP 1060/50000, loss = 3.093398 (0.085 sec/batch), lr: 0.003000
-2025-12-03 15:53:25 INFO: Finished STEP 1080/50000, loss = 3.780082 (0.088 sec/batch), lr: 0.003000
-2025-12-03 15:53:27 INFO: Finished STEP 1100/50000, loss = 3.758047 (0.075 sec/batch), lr: 0.003000
-2025-12-03 15:53:27 INFO: Evaluating on dev set...
-2025-12-03 15:53:27 INFO: LAS	MLAS	BLEX
-2025-12-03 15:53:27 INFO: 59.41	49.05	52.85
-2025-12-03 15:53:27 INFO: step 1100: train_loss = 3.779099, dev_score = 0.5941
-2025-12-03 15:53:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:53:28 INFO: new model checkpoint saved.
-2025-12-03 15:53:30 INFO: Finished STEP 1120/50000, loss = 3.774802 (0.088 sec/batch), lr: 0.003000
-2025-12-03 15:53:31 INFO: Finished STEP 1140/50000, loss = 2.977976 (0.089 sec/batch), lr: 0.003000
-2025-12-03 15:53:33 INFO: Finished STEP 1160/50000, loss = 3.838738 (0.078 sec/batch), lr: 0.003000
-2025-12-03 15:53:35 INFO: Finished STEP 1180/50000, loss = 4.660880 (0.080 sec/batch), lr: 0.003000
-2025-12-03 15:53:36 INFO: Finished STEP 1200/50000, loss = 2.489745 (0.088 sec/batch), lr: 0.003000
-2025-12-03 15:53:36 INFO: Evaluating on dev set...
-2025-12-03 15:53:37 INFO: LAS	MLAS	BLEX
-2025-12-03 15:53:37 INFO: 58.66	48.31	51.69
-2025-12-03 15:53:37 INFO: step 1200: train_loss = 3.720397, dev_score = 0.5866
-2025-12-03 15:53:37 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:53:37 INFO: new model checkpoint saved.
-2025-12-03 15:53:39 INFO: Finished STEP 1220/50000, loss = 3.247226 (0.090 sec/batch), lr: 0.003000
-2025-12-03 15:53:41 INFO: Finished STEP 1240/50000, loss = 4.093720 (0.079 sec/batch), lr: 0.003000
-2025-12-03 15:53:43 INFO: Finished STEP 1260/50000, loss = 2.970687 (0.080 sec/batch), lr: 0.003000
-2025-12-03 15:53:44 INFO: Finished STEP 1280/50000, loss = 3.959898 (0.087 sec/batch), lr: 0.003000
-2025-12-03 15:53:46 INFO: Finished STEP 1300/50000, loss = 3.890834 (0.081 sec/batch), lr: 0.003000
-2025-12-03 15:53:46 INFO: Evaluating on dev set...
-2025-12-03 15:53:47 INFO: LAS	MLAS	BLEX
-2025-12-03 15:53:47 INFO: 57.67	50.00	52.97
-2025-12-03 15:53:47 INFO: step 1300: train_loss = 3.945667, dev_score = 0.5767
-2025-12-03 15:53:47 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:53:47 INFO: new model checkpoint saved.
-2025-12-03 15:53:49 INFO: Finished STEP 1320/50000, loss = 4.299366 (0.076 sec/batch), lr: 0.003000
-2025-12-03 15:53:51 INFO: Finished STEP 1340/50000, loss = 3.080094 (0.082 sec/batch), lr: 0.003000
-2025-12-03 15:53:52 INFO: Finished STEP 1360/50000, loss = 5.326971 (0.087 sec/batch), lr: 0.003000
-2025-12-03 15:53:54 INFO: Finished STEP 1380/50000, loss = 4.148820 (0.083 sec/batch), lr: 0.003000
-2025-12-03 15:53:56 INFO: Finished STEP 1400/50000, loss = 3.131835 (0.085 sec/batch), lr: 0.003000
-2025-12-03 15:53:56 INFO: Evaluating on dev set...
-2025-12-03 15:53:56 INFO: LAS	MLAS	BLEX
-2025-12-03 15:53:56 INFO: 59.16	49.47	52.85
-2025-12-03 15:53:56 INFO: step 1400: train_loss = 3.828261, dev_score = 0.5916
-2025-12-03 15:53:57 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:53:57 INFO: new model checkpoint saved.
-2025-12-03 15:53:58 INFO: Finished STEP 1420/50000, loss = 3.877927 (0.079 sec/batch), lr: 0.003000
-2025-12-03 15:54:00 INFO: Finished STEP 1440/50000, loss = 2.828221 (0.077 sec/batch), lr: 0.003000
-2025-12-03 15:54:02 INFO: Finished STEP 1460/50000, loss = 2.771763 (0.088 sec/batch), lr: 0.003000
-2025-12-03 15:54:03 INFO: Finished STEP 1480/50000, loss = 3.887072 (0.082 sec/batch), lr: 0.003000
-2025-12-03 15:54:05 INFO: Finished STEP 1500/50000, loss = 3.089223 (0.080 sec/batch), lr: 0.003000
-2025-12-03 15:54:05 INFO: Evaluating on dev set...
-2025-12-03 15:54:05 INFO: LAS	MLAS	BLEX
-2025-12-03 15:54:05 INFO: 60.89	49.47	54.12
-2025-12-03 15:54:05 INFO: step 1500: train_loss = 3.662534, dev_score = 0.6089
-2025-12-03 15:54:06 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 15:54:06 INFO: new best model saved.
-2025-12-03 15:54:06 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:54:06 INFO: new model checkpoint saved.
-2025-12-03 15:54:08 INFO: Finished STEP 1520/50000, loss = 3.424534 (0.079 sec/batch), lr: 0.003000
-2025-12-03 15:54:10 INFO: Finished STEP 1540/50000, loss = 4.257630 (0.081 sec/batch), lr: 0.003000
-2025-12-03 15:54:11 INFO: Finished STEP 1560/50000, loss = 4.004986 (0.078 sec/batch), lr: 0.003000
-2025-12-03 15:54:13 INFO: Finished STEP 1580/50000, loss = 4.259016 (0.076 sec/batch), lr: 0.003000
-2025-12-03 15:54:14 INFO: Finished STEP 1600/50000, loss = 4.886831 (0.071 sec/batch), lr: 0.003000
-2025-12-03 15:54:14 INFO: Evaluating on dev set...
-2025-12-03 15:54:15 INFO: LAS	MLAS	BLEX
-2025-12-03 15:54:15 INFO: 58.66	49.47	52.01
-2025-12-03 15:54:15 INFO: step 1600: train_loss = 3.940190, dev_score = 0.5866
-2025-12-03 15:54:15 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:54:15 INFO: new model checkpoint saved.
-2025-12-03 15:54:17 INFO: Finished STEP 1620/50000, loss = 3.486598 (0.072 sec/batch), lr: 0.003000
-2025-12-03 15:54:19 INFO: Finished STEP 1640/50000, loss = 3.426835 (0.076 sec/batch), lr: 0.003000
-2025-12-03 15:54:20 INFO: Finished STEP 1660/50000, loss = 3.082500 (0.077 sec/batch), lr: 0.003000
-2025-12-03 15:54:22 INFO: Finished STEP 1680/50000, loss = 2.587858 (0.080 sec/batch), lr: 0.003000
-2025-12-03 15:54:23 INFO: Finished STEP 1700/50000, loss = 4.049259 (0.081 sec/batch), lr: 0.003000
-2025-12-03 15:54:23 INFO: Evaluating on dev set...
-2025-12-03 15:54:24 INFO: LAS	MLAS	BLEX
-2025-12-03 15:54:24 INFO: 59.41	50.63	54.43
-2025-12-03 15:54:24 INFO: step 1700: train_loss = 3.731144, dev_score = 0.5941
-2025-12-03 15:54:24 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:54:24 INFO: new model checkpoint saved.
-2025-12-03 15:54:26 INFO: Finished STEP 1720/50000, loss = 2.259671 (0.076 sec/batch), lr: 0.003000
-2025-12-03 15:54:28 INFO: Finished STEP 1740/50000, loss = 5.103413 (0.079 sec/batch), lr: 0.003000
-2025-12-03 15:54:29 INFO: Finished STEP 1760/50000, loss = 3.651329 (0.072 sec/batch), lr: 0.003000
-2025-12-03 15:54:31 INFO: Finished STEP 1780/50000, loss = 3.879523 (0.078 sec/batch), lr: 0.003000
-2025-12-03 15:54:32 INFO: Finished STEP 1800/50000, loss = 3.246203 (0.078 sec/batch), lr: 0.003000
-2025-12-03 15:54:32 INFO: Evaluating on dev set...
-2025-12-03 15:54:33 INFO: LAS	MLAS	BLEX
-2025-12-03 15:54:33 INFO: 60.89	53.39	55.93
-2025-12-03 15:54:33 INFO: step 1800: train_loss = 3.826308, dev_score = 0.6089
-2025-12-03 15:54:33 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 15:54:33 INFO: new best model saved.
-2025-12-03 15:54:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:54:34 INFO: new model checkpoint saved.
-2025-12-03 15:54:35 INFO: Finished STEP 1820/50000, loss = 4.498244 (0.072 sec/batch), lr: 0.003000
-2025-12-03 15:54:37 INFO: Finished STEP 1840/50000, loss = 3.589321 (0.078 sec/batch), lr: 0.003000
-2025-12-03 15:54:39 INFO: Finished STEP 1860/50000, loss = 2.884829 (0.077 sec/batch), lr: 0.003000
-2025-12-03 15:54:40 INFO: Finished STEP 1880/50000, loss = 3.823880 (0.082 sec/batch), lr: 0.003000
-2025-12-03 15:54:42 INFO: Finished STEP 1900/50000, loss = 2.696285 (0.082 sec/batch), lr: 0.003000
-2025-12-03 15:54:42 INFO: Evaluating on dev set...
-2025-12-03 15:54:42 INFO: LAS	MLAS	BLEX
-2025-12-03 15:54:42 INFO: 60.64	51.16	55.39
-2025-12-03 15:54:42 INFO: step 1900: train_loss = 4.029873, dev_score = 0.6064
-2025-12-03 15:54:43 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:54:43 INFO: new model checkpoint saved.
-2025-12-03 15:54:44 INFO: Finished STEP 1920/50000, loss = 5.128090 (0.068 sec/batch), lr: 0.003000
-2025-12-03 15:54:46 INFO: Finished STEP 1940/50000, loss = 3.145783 (0.078 sec/batch), lr: 0.003000
-2025-12-03 15:54:47 INFO: Finished STEP 1960/50000, loss = 3.195472 (0.075 sec/batch), lr: 0.003000
-2025-12-03 15:54:49 INFO: Finished STEP 1980/50000, loss = 3.154953 (0.072 sec/batch), lr: 0.003000
-2025-12-03 15:54:51 INFO: Finished STEP 2000/50000, loss = 3.763106 (0.080 sec/batch), lr: 0.003000
-2025-12-03 15:54:51 INFO: Evaluating on dev set...
-2025-12-03 15:54:51 INFO: LAS	MLAS	BLEX
-2025-12-03 15:54:51 INFO: 61.39	50.85	54.66
-2025-12-03 15:54:51 INFO: step 2000: train_loss = 3.764299, dev_score = 0.6139
-2025-12-03 15:54:51 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 15:54:51 INFO: new best model saved.
-2025-12-03 15:54:52 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:54:52 INFO: new model checkpoint saved.
-2025-12-03 15:54:53 INFO: Finished STEP 2020/50000, loss = 4.129125 (0.077 sec/batch), lr: 0.003000
-2025-12-03 15:54:55 INFO: Finished STEP 2040/50000, loss = 4.000008 (0.076 sec/batch), lr: 0.003000
-2025-12-03 15:54:56 INFO: Finished STEP 2060/50000, loss = 3.594474 (0.076 sec/batch), lr: 0.003000
-2025-12-03 15:54:58 INFO: Finished STEP 2080/50000, loss = 5.156559 (0.070 sec/batch), lr: 0.003000
-2025-12-03 15:54:59 INFO: Finished STEP 2100/50000, loss = 2.780583 (0.076 sec/batch), lr: 0.003000
-2025-12-03 15:54:59 INFO: Evaluating on dev set...
-2025-12-03 15:55:00 INFO: LAS	MLAS	BLEX
-2025-12-03 15:55:00 INFO: 58.91	50.32	53.28
-2025-12-03 15:55:00 INFO: step 2100: train_loss = 3.970306, dev_score = 0.5891
-2025-12-03 15:55:01 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:55:01 INFO: new model checkpoint saved.
-2025-12-03 15:55:02 INFO: Finished STEP 2120/50000, loss = 3.461082 (0.078 sec/batch), lr: 0.003000
-2025-12-03 15:55:04 INFO: Finished STEP 2140/50000, loss = 4.371090 (0.068 sec/batch), lr: 0.003000
-2025-12-03 15:55:05 INFO: Finished STEP 2160/50000, loss = 3.734883 (0.074 sec/batch), lr: 0.003000
-2025-12-03 15:55:07 INFO: Finished STEP 2180/50000, loss = 2.186298 (0.077 sec/batch), lr: 0.003000
-2025-12-03 15:55:08 INFO: Finished STEP 2200/50000, loss = 3.871332 (0.078 sec/batch), lr: 0.003000
-2025-12-03 15:55:08 INFO: Evaluating on dev set...
-2025-12-03 15:55:09 INFO: LAS	MLAS	BLEX
-2025-12-03 15:55:09 INFO: 58.66	48.73	52.97
-2025-12-03 15:55:09 INFO: step 2200: train_loss = 3.926589, dev_score = 0.5866
-2025-12-03 15:55:09 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:55:09 INFO: new model checkpoint saved.
-2025-12-03 15:55:11 INFO: Finished STEP 2220/50000, loss = 2.865967 (0.073 sec/batch), lr: 0.003000
-2025-12-03 15:55:12 INFO: Finished STEP 2240/50000, loss = 2.578629 (0.069 sec/batch), lr: 0.003000
-2025-12-03 15:55:14 INFO: Finished STEP 2260/50000, loss = 3.295215 (0.068 sec/batch), lr: 0.003000
-2025-12-03 15:55:15 INFO: Finished STEP 2280/50000, loss = 3.903890 (0.074 sec/batch), lr: 0.003000
-2025-12-03 15:55:17 INFO: Finished STEP 2300/50000, loss = 3.614969 (0.071 sec/batch), lr: 0.003000
-2025-12-03 15:55:17 INFO: Evaluating on dev set...
-2025-12-03 15:55:17 INFO: LAS	MLAS	BLEX
-2025-12-03 15:55:17 INFO: 60.40	52.43	56.66
-2025-12-03 15:55:17 INFO: step 2300: train_loss = 3.953848, dev_score = 0.6040
-2025-12-03 15:55:18 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:55:18 INFO: new model checkpoint saved.
-2025-12-03 15:55:19 INFO: Finished STEP 2320/50000, loss = 3.836884 (0.076 sec/batch), lr: 0.003000
-2025-12-03 15:55:21 INFO: Finished STEP 2340/50000, loss = 2.887274 (0.079 sec/batch), lr: 0.003000
-2025-12-03 15:55:22 INFO: Finished STEP 2360/50000, loss = 4.795152 (0.079 sec/batch), lr: 0.003000
-2025-12-03 15:55:24 INFO: Finished STEP 2380/50000, loss = 4.483939 (0.074 sec/batch), lr: 0.003000
-2025-12-03 15:55:25 INFO: Finished STEP 2400/50000, loss = 4.851534 (0.075 sec/batch), lr: 0.003000
-2025-12-03 15:55:25 INFO: Evaluating on dev set...
-2025-12-03 15:55:26 INFO: LAS	MLAS	BLEX
-2025-12-03 15:55:26 INFO: 58.91	50.21	54.85
-2025-12-03 15:55:26 INFO: step 2400: train_loss = 3.932371, dev_score = 0.5891
-2025-12-03 15:55:27 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:55:27 INFO: new model checkpoint saved.
-2025-12-03 15:55:28 INFO: Finished STEP 2420/50000, loss = 4.692697 (0.067 sec/batch), lr: 0.003000
-2025-12-03 15:55:30 INFO: Finished STEP 2440/50000, loss = 2.805047 (0.068 sec/batch), lr: 0.003000
-2025-12-03 15:55:31 INFO: Finished STEP 2460/50000, loss = 3.715670 (0.077 sec/batch), lr: 0.003000
-2025-12-03 15:55:32 INFO: Finished STEP 2480/50000, loss = 5.003852 (0.072 sec/batch), lr: 0.003000
-2025-12-03 15:55:34 INFO: Finished STEP 2500/50000, loss = 5.362862 (0.071 sec/batch), lr: 0.003000
-2025-12-03 15:55:34 INFO: Evaluating on dev set...
-2025-12-03 15:55:34 INFO: LAS	MLAS	BLEX
-2025-12-03 15:55:34 INFO: 62.87	54.12	57.51
-2025-12-03 15:55:34 INFO: step 2500: train_loss = 3.716296, dev_score = 0.6287
-2025-12-03 15:55:35 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 15:55:35 INFO: new best model saved.
-2025-12-03 15:55:35 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:55:35 INFO: new model checkpoint saved.
-2025-12-03 15:55:37 INFO: Finished STEP 2520/50000, loss = 4.285573 (0.073 sec/batch), lr: 0.003000
-2025-12-03 15:55:38 INFO: Finished STEP 2540/50000, loss = 3.220155 (0.071 sec/batch), lr: 0.003000
-2025-12-03 15:55:40 INFO: Finished STEP 2560/50000, loss = 4.155645 (0.077 sec/batch), lr: 0.003000
-2025-12-03 15:55:41 INFO: Finished STEP 2580/50000, loss = 3.434030 (0.073 sec/batch), lr: 0.003000
-2025-12-03 15:55:43 INFO: Finished STEP 2600/50000, loss = 2.680579 (0.076 sec/batch), lr: 0.003000
-2025-12-03 15:55:43 INFO: Evaluating on dev set...
-2025-12-03 15:55:43 INFO: LAS	MLAS	BLEX
-2025-12-03 15:55:43 INFO: 62.87	54.66	58.05
-2025-12-03 15:55:43 INFO: step 2600: train_loss = 4.018973, dev_score = 0.6287
-2025-12-03 15:55:43 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 15:55:43 INFO: new best model saved.
-2025-12-03 15:55:44 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:55:44 INFO: new model checkpoint saved.
-2025-12-03 15:55:45 INFO: Finished STEP 2620/50000, loss = 2.524696 (0.072 sec/batch), lr: 0.003000
-2025-12-03 15:55:47 INFO: Finished STEP 2640/50000, loss = 3.829545 (0.065 sec/batch), lr: 0.003000
-2025-12-03 15:55:48 INFO: Finished STEP 2660/50000, loss = 3.246467 (0.069 sec/batch), lr: 0.003000
-2025-12-03 15:55:50 INFO: Finished STEP 2680/50000, loss = 2.396587 (0.068 sec/batch), lr: 0.003000
-2025-12-03 15:55:51 INFO: Finished STEP 2700/50000, loss = 3.570580 (0.071 sec/batch), lr: 0.003000
-2025-12-03 15:55:51 INFO: Evaluating on dev set...
-2025-12-03 15:55:52 INFO: LAS	MLAS	BLEX
-2025-12-03 15:55:52 INFO: 58.91	49.58	54.24
-2025-12-03 15:55:52 INFO: step 2700: train_loss = 3.918322, dev_score = 0.5891
-2025-12-03 15:55:52 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:55:52 INFO: new model checkpoint saved.
-2025-12-03 15:55:54 INFO: Finished STEP 2720/50000, loss = 2.712030 (0.072 sec/batch), lr: 0.003000
-2025-12-03 15:55:55 INFO: Finished STEP 2740/50000, loss = 5.811014 (0.077 sec/batch), lr: 0.003000
-2025-12-03 15:55:57 INFO: Finished STEP 2760/50000, loss = 3.773255 (0.076 sec/batch), lr: 0.003000
-2025-12-03 15:55:58 INFO: Finished STEP 2780/50000, loss = 2.087427 (0.075 sec/batch), lr: 0.003000
-2025-12-03 15:56:00 INFO: Finished STEP 2800/50000, loss = 3.811720 (0.063 sec/batch), lr: 0.003000
-2025-12-03 15:56:00 INFO: Evaluating on dev set...
-2025-12-03 15:56:00 INFO: LAS	MLAS	BLEX
-2025-12-03 15:56:00 INFO: 61.63	51.59	55.39
-2025-12-03 15:56:00 INFO: step 2800: train_loss = 3.902705, dev_score = 0.6163
-2025-12-03 15:56:01 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:56:01 INFO: new model checkpoint saved.
-2025-12-03 15:56:02 INFO: Finished STEP 2820/50000, loss = 3.443505 (0.070 sec/batch), lr: 0.003000
-2025-12-03 15:56:04 INFO: Finished STEP 2840/50000, loss = 3.684120 (0.064 sec/batch), lr: 0.003000
-2025-12-03 15:56:05 INFO: Finished STEP 2860/50000, loss = 5.325919 (0.071 sec/batch), lr: 0.003000
-2025-12-03 15:56:06 INFO: Finished STEP 2880/50000, loss = 3.516696 (0.068 sec/batch), lr: 0.003000
-2025-12-03 15:56:08 INFO: Finished STEP 2900/50000, loss = 3.573216 (0.072 sec/batch), lr: 0.003000
-2025-12-03 15:56:08 INFO: Evaluating on dev set...
-2025-12-03 15:56:09 INFO: LAS	MLAS	BLEX
-2025-12-03 15:56:09 INFO: 58.17	48.20	53.28
-2025-12-03 15:56:09 INFO: step 2900: train_loss = 4.154457, dev_score = 0.5817
-2025-12-03 15:56:09 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:56:09 INFO: new model checkpoint saved.
-2025-12-03 15:56:11 INFO: Finished STEP 2920/50000, loss = 2.522206 (0.077 sec/batch), lr: 0.003000
-2025-12-03 15:56:12 INFO: Finished STEP 2940/50000, loss = 2.687997 (0.082 sec/batch), lr: 0.003000
-2025-12-03 15:56:13 INFO: Finished STEP 2960/50000, loss = 3.291323 (0.078 sec/batch), lr: 0.003000
-2025-12-03 15:56:15 INFO: Finished STEP 2980/50000, loss = 4.553081 (0.069 sec/batch), lr: 0.003000
-2025-12-03 15:56:16 INFO: Finished STEP 3000/50000, loss = 3.136598 (0.071 sec/batch), lr: 0.003000
-2025-12-03 15:56:16 INFO: Evaluating on dev set...
-2025-12-03 15:56:17 INFO: LAS	MLAS	BLEX
-2025-12-03 15:56:17 INFO: 61.14	52.43	55.39
-2025-12-03 15:56:17 INFO: step 3000: train_loss = 3.949049, dev_score = 0.6114
-2025-12-03 15:56:17 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:56:17 INFO: new model checkpoint saved.
-2025-12-03 15:56:19 INFO: Finished STEP 3020/50000, loss = 3.020242 (0.077 sec/batch), lr: 0.003000
-2025-12-03 15:56:20 INFO: Finished STEP 3040/50000, loss = 4.876935 (0.069 sec/batch), lr: 0.003000
-2025-12-03 15:56:22 INFO: Finished STEP 3060/50000, loss = 3.937078 (0.080 sec/batch), lr: 0.003000
-2025-12-03 15:56:23 INFO: Finished STEP 3080/50000, loss = 4.811732 (0.064 sec/batch), lr: 0.003000
-2025-12-03 15:56:25 INFO: Finished STEP 3100/50000, loss = 3.313049 (0.073 sec/batch), lr: 0.003000
-2025-12-03 15:56:25 INFO: Evaluating on dev set...
-2025-12-03 15:56:25 INFO: LAS	MLAS	BLEX
-2025-12-03 15:56:25 INFO: 62.13	54.12	58.35
-2025-12-03 15:56:25 INFO: step 3100: train_loss = 4.106587, dev_score = 0.6213
-2025-12-03 15:56:26 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:56:26 INFO: new model checkpoint saved.
-2025-12-03 15:56:27 INFO: Finished STEP 3120/50000, loss = 3.817453 (0.067 sec/batch), lr: 0.003000
-2025-12-03 15:56:29 INFO: Finished STEP 3140/50000, loss = 5.620703 (0.070 sec/batch), lr: 0.003000
-2025-12-03 15:56:30 INFO: Finished STEP 3160/50000, loss = 4.408039 (0.078 sec/batch), lr: 0.003000
-2025-12-03 15:56:32 INFO: Finished STEP 3180/50000, loss = 2.225734 (0.074 sec/batch), lr: 0.003000
-2025-12-03 15:56:33 INFO: Finished STEP 3200/50000, loss = 6.085883 (0.082 sec/batch), lr: 0.003000
-2025-12-03 15:56:33 INFO: Evaluating on dev set...
-2025-12-03 15:56:34 INFO: LAS	MLAS	BLEX
-2025-12-03 15:56:34 INFO: 58.91	51.59	53.70
-2025-12-03 15:56:34 INFO: step 3200: train_loss = 4.045106, dev_score = 0.5891
-2025-12-03 15:56:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:56:34 INFO: new model checkpoint saved.
-2025-12-03 15:56:36 INFO: Finished STEP 3220/50000, loss = 3.643058 (0.064 sec/batch), lr: 0.003000
-2025-12-03 15:56:37 INFO: Finished STEP 3240/50000, loss = 3.506227 (0.068 sec/batch), lr: 0.003000
-2025-12-03 15:56:38 INFO: Finished STEP 3260/50000, loss = 2.930928 (0.072 sec/batch), lr: 0.003000
-2025-12-03 15:56:40 INFO: Finished STEP 3280/50000, loss = 4.896526 (0.071 sec/batch), lr: 0.003000
-2025-12-03 15:56:41 INFO: Finished STEP 3300/50000, loss = 4.648397 (0.077 sec/batch), lr: 0.003000
-2025-12-03 15:56:41 INFO: Evaluating on dev set...
-2025-12-03 15:56:42 INFO: LAS	MLAS	BLEX
-2025-12-03 15:56:42 INFO: 59.16	52.01	52.43
-2025-12-03 15:56:42 INFO: step 3300: train_loss = 3.880126, dev_score = 0.5916
-2025-12-03 15:56:42 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:56:42 INFO: new model checkpoint saved.
-2025-12-03 15:56:44 INFO: Finished STEP 3320/50000, loss = 4.656130 (0.065 sec/batch), lr: 0.003000
-2025-12-03 15:56:45 INFO: Finished STEP 3340/50000, loss = 4.644033 (0.070 sec/batch), lr: 0.003000
-2025-12-03 15:56:47 INFO: Finished STEP 3360/50000, loss = 4.342103 (0.070 sec/batch), lr: 0.003000
-2025-12-03 15:56:48 INFO: Finished STEP 3380/50000, loss = 1.757741 (0.069 sec/batch), lr: 0.003000
-2025-12-03 15:56:49 INFO: Finished STEP 3400/50000, loss = 3.296278 (0.077 sec/batch), lr: 0.003000
-2025-12-03 15:56:49 INFO: Evaluating on dev set...
-2025-12-03 15:56:50 INFO: LAS	MLAS	BLEX
-2025-12-03 15:56:50 INFO: 57.92	48.63	52.43
-2025-12-03 15:56:50 INFO: step 3400: train_loss = 3.761222, dev_score = 0.5792
-2025-12-03 15:56:51 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:56:51 INFO: new model checkpoint saved.
-2025-12-03 15:56:52 INFO: Finished STEP 3420/50000, loss = 4.175681 (0.063 sec/batch), lr: 0.003000
-2025-12-03 15:56:53 INFO: Finished STEP 3440/50000, loss = 3.109319 (0.072 sec/batch), lr: 0.003000
-2025-12-03 15:56:55 INFO: Finished STEP 3460/50000, loss = 5.698625 (0.070 sec/batch), lr: 0.003000
-2025-12-03 15:56:56 INFO: Finished STEP 3480/50000, loss = 4.906111 (0.069 sec/batch), lr: 0.003000
-2025-12-03 15:56:58 INFO: Finished STEP 3500/50000, loss = 2.886077 (0.068 sec/batch), lr: 0.003000
-2025-12-03 15:56:58 INFO: Evaluating on dev set...
-2025-12-03 15:56:58 INFO: LAS	MLAS	BLEX
-2025-12-03 15:56:58 INFO: 60.64	51.48	55.70
-2025-12-03 15:56:58 INFO: step 3500: train_loss = 3.804499, dev_score = 0.6064
-2025-12-03 15:56:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
-2025-12-03 15:56:59 INFO: new model checkpoint saved.
-2025-12-03 15:57:00 INFO: Finished STEP 3520/50000, loss = 2.301594 (0.065 sec/batch), lr: 0.003000
-2025-12-03 15:57:01 INFO: Finished STEP 3540/50000, loss = 4.423633 (0.075 sec/batch), lr: 0.003000
-2025-12-03 15:57:03 INFO: Finished STEP 3560/50000, loss = 3.613044 (0.063 sec/batch), lr: 0.003000
-2025-12-03 15:57:04 INFO: Finished STEP 3580/50000, loss = 3.496284 (0.067 sec/batch), lr: 0.003000
-2025-12-03 15:57:06 INFO: Finished STEP 3600/50000, loss = 6.527389 (0.066 sec/batch), lr: 0.003000
-2025-12-03 15:57:06 INFO: Evaluating on dev set...
-2025-12-03 15:57:06 INFO: LAS	MLAS	BLEX
-2025-12-03 15:57:06 INFO: 60.64	51.27	55.08
-2025-12-03 15:57:06 INFO: step 3600: train_loss = 4.020496, dev_score = 0.6064
-2025-12-03 15:57:06 INFO: Training ended with 3600 steps.
-2025-12-03 15:57:06 INFO: Best dev F1 = 62.87, at iteration = 2500
-2025-12-03 15:57:07 INFO: Running dev depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
-2025-12-03 15:57:07 INFO: Running parser in predict mode
-2025-12-03 15:57:07 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 15:57:09 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 15:57:09 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 15:57:09 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 15:57:09 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 15:57:10 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 15:57:10 INFO: Loading data with batch size 32...
-2025-12-03 15:57:10 DEBUG: 9 batches created.
-2025-12-03 15:57:10 INFO: F1 scores for each dependency:
-  Note that unlabeled attachment errors hurt the labeled attachment scores
-       acl: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
- acl:relcl: p 0.2222 r 0.2857 f1 0.2500 (7 actual)
-     advcl: p 0.1250 r 0.2000 f1 0.1538 (5 actual)
-    advmod: p 0.5357 r 0.6000 f1 0.5660 (25 actual)
-      amod: p 0.9231 r 0.7742 f1 0.8421 (31 actual)
-     appos: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-       aux: p 0.8182 r 0.8182 f1 0.8182 (11 actual)
-      case: p 0.9623 r 0.9107 f1 0.9358 (56 actual)
-        cc: p 0.6429 r 0.6923 f1 0.6667 (13 actual)
-     ccomp: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-      conj: p 0.2667 r 0.3333 f1 0.2963 (12 actual)
-       cop: p 0.5000 r 0.3333 f1 0.4000 (3 actual)
-     csubj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-       det: p 0.8696 r 0.9091 f1 0.8889 (22 actual)
-      expl: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-      iobj: p 0.5000 r 0.5000 f1 0.5000 (2 actual)
-      mark: p 0.4167 r 0.4167 f1 0.4167 (12 actual)
-      nmod: p 0.4000 r 0.5333 f1 0.4571 (15 actual)
- nmod:poss: p 1.0000 r 0.8947 f1 0.9444 (19 actual)
-     nsubj: p 0.5417 r 0.7647 f1 0.6341 (17 actual)
-nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
-       obj: p 0.6071 r 0.7727 f1 0.6800 (22 actual)
-       obl: p 0.6154 r 0.5854 f1 0.6000 (41 actual)
- obl:agent: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-    orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
- parataxis: p 0.1111 r 0.3333 f1 0.1667 (3 actual)
-     punct: p 0.4231 r 0.4231 f1 0.4231 (52 actual)
-      root: p 0.4444 r 0.4444 f1 0.4444 (9 actual)
-     xcomp: p 0.5000 r 0.1250 f1 0.2000 (8 actual)
-2025-12-03 15:57:10 INFO: LAS	MLAS	BLEX
-2025-12-03 15:57:10 INFO: 62.87	54.66	58.05
-2025-12-03 15:57:10 INFO: Parser score:
-2025-12-03 15:57:10 INFO: sv_diachronic 62.87
-2025-12-03 15:57:10 INFO: Finished running dev set on
-UD_Swedish-diachronic
-  UAS   LAS  CLAS  MLAS  BLEX
-68.56 62.87 58.05 54.66 58.05
-2025-12-03 15:57:10 INFO: Running test depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
-2025-12-03 15:57:10 INFO: Running parser in predict mode
-2025-12-03 15:57:10 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
-2025-12-03 15:57:12 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
-2025-12-03 15:57:12 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 15:57:12 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
-2025-12-03 15:57:12 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
-2025-12-03 15:57:12 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
-2025-12-03 15:57:12 INFO: Loading data with batch size 32...
-2025-12-03 15:57:12 DEBUG: 93 batches created.
-2025-12-03 15:57:17 INFO: F1 scores for each dependency:
-  Note that unlabeled attachment errors hurt the labeled attachment scores
-         acl: p 0.2778 r 0.1562 f1 0.2000 (32 actual)
-   acl:cleft: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-   acl:relcl: p 0.2075 r 0.1467 f1 0.1719 (75 actual)
-       advcl: p 0.0842 r 0.1333 f1 0.1032 (60 actual)
- advcl:relcl: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-      advmod: p 0.5520 r 0.5746 f1 0.5631 (268 actual)
-        amod: p 0.8596 r 0.8522 f1 0.8559 (230 actual)
-       appos: p 0.0000 r 0.0000 f1 0.0000 (13 actual)
-         aux: p 0.8734 r 0.8214 f1 0.8466 (84 actual)
-    aux:pass: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-        case: p 0.8481 r 0.8231 f1 0.8354 (373 actual)
-          cc: p 0.5833 r 0.5871 f1 0.5852 (155 actual)
-       ccomp: p 0.0000 r 0.0000 f1 0.0000 (35 actual)
-compound:prt: p 0.7727 r 0.8095 f1 0.7907 (21 actual)
-        conj: p 0.2397 r 0.2215 f1 0.2303 (158 actual)
-         cop: p 0.7714 r 0.5870 f1 0.6667 (46 actual)
-       csubj: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-         dep: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-         det: p 0.8960 r 0.8702 f1 0.8829 (208 actual)
-   discourse: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
-  dislocated: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-        expl: p 0.0000 r 0.0000 f1 0.0000 (11 actual)
-     expl:pv: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-       fixed: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
-        flat: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-   flat:name: p 0.0000 r 0.0000 f1 0.0000 (12 actual)
-    goeswith: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
-        iobj: p 0.3333 r 0.0714 f1 0.1176 (14 actual)
-        mark: p 0.6757 r 0.6536 f1 0.6645 (153 actual)
-        nmod: p 0.3798 r 0.4804 f1 0.4242 (102 actual)
-   nmod:poss: p 0.8707 r 0.9014 f1 0.8858 (142 actual)
-       nsubj: p 0.5673 r 0.6321 f1 0.5980 (280 actual)
-  nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (25 actual)
-      nummod: p 0.7500 r 0.9000 f1 0.8182 (10 actual)
-         obj: p 0.5628 r 0.7104 f1 0.6280 (183 actual)
-         obl: p 0.4857 r 0.5504 f1 0.5160 (278 actual)
-   obl:agent: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
-      orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-   parataxis: p 0.0333 r 0.1667 f1 0.0556 (18 actual)
-       punct: p 0.4262 r 0.4282 f1 0.4272 (425 actual)
-  reparandum: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
-        root: p 0.5354 r 0.5354 f1 0.5354 (99 actual)
-    vocative: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
-       xcomp: p 0.5750 r 0.3067 f1 0.4000 (75 actual)
-2025-12-03 15:57:17 INFO: LAS	MLAS	BLEX
-2025-12-03 15:57:17 INFO: 58.54	50.12	53.03
-2025-12-03 15:57:17 INFO: Parser score:
-2025-12-03 15:57:17 INFO: sv_diachronic 58.54
-2025-12-03 15:57:17 INFO: Finished running test set on
-UD_Swedish-diachronic
-  UAS   LAS  CLAS  MLAS  BLEX
-67.16 58.54 53.03 50.12 53.03
-DONE.
-Full log saved to: logs/log_conll17.pt_sv_diachron_de_lit_20251203_154509.txt
-Symlink updated: logs/latest.txt → log_conll17.pt_sv_diachron_de_lit_20251203_154509.txt