albertdatzira

qwen3-8b-htr-vines-lora

Key Result

Table with columns: Approach, CER, Pages worsened
Approach	CER	Pages worsened
HTR only (Balakirev)	10.86%	—
Llama 3.1 70B (general, no fine-tuning)	21.20%	all
This model + fallback	7.80%	0

The fine-tuned 8B model outperforms a general 70B model by 2.7× — demonstrating that domain-specific fine-tuning is essential for historical manuscript correction.

Model Details

Base model: Qwen/Qwen3-8B
Method: LoRA (Low-Rank Adaptation)
LoRA rank: 16, alpha: 32, dropout: 0.05
Trainable parameters: 43.6M (0.53% of total)
Training data: 583 page-level pairs (HTR raw → ground truth)
Training time: 56 minutes on NVIDIA DGX Spark
Epochs: 3
Language: Spanish (primary), French, Catalan

Usage

python

import torchfrom peft import PeftModelfrom transformers import AutoModelForCausalLM, AutoTokenizer # Load base model + LoRA adaptersbase_model = AutoModelForCausalLM.from_pretrained(    "Qwen/Qwen3-8B",    trust_remote_code=True,    device_map="auto",    torch_dtype=torch.bfloat16,)model = PeftModel.from_pretrained(base_model, "albertdatzira/qwen3-8b-htr-vines-lora")tokenizer = AutoTokenizer.from_pretrained("albertdatzira/qwen3-8b-htr-vines-lora") model.eval() # Correct HTR textsystem_prompt = (    "Ets un expert en post-correcció d'HTR dels diaris de Ramon Viñes. "    "Corregeix els errors de reconeixement del text manuscrit. "    "Només corregeix errors d'HTR evidents. No modifiquis l'ortografia "    "de l'autor, noms propis, ni símbols especials (¶, ¬, ⟦ ⟧).") htr_text = "para el almuerzo (á mediddia). Me hristes son todos estos aniversarios" messages = [    {"role": "system", "content": system_prompt},    {"role": "user", "content": f"Corregeix el següent text HTR:\n{htr_text}"},] text = tokenizer.apply_chat_template(    messages, tokenize=False, add_generation_prompt=True,    enable_thinking=False,)inputs = tokenizer(text, return_tensors="pt").to(model.device) with torch.no_grad():    outputs = model.generate(        **inputs, max_new_tokens=4096, do_sample=False,        pad_token_id=tokenizer.pad_token_id,    ) generated = outputs[0][inputs["input_ids"].shape[-1]:]corrected = tokenizer.decode(generated, skip_special_tokens=True)print(corrected)# -> "para el almuerzo (á mediodia). Qué tristes son todos estos aniversarios"

Training Data

The model was trained on 583 page-level pairs from the RV-training repository:

Input: raw HTR output from the Balakirev model (Kraken/eScriptorium)
Output: human-validated ground truth transcriptions
Split: by year (not random) to prevent data leakage
Filtering: pages with GT < 50 chars or CER > 80% excluded

Limitations

Best on easy pages: the model achieves 89% page improvement rate on test set pages (CER ~10-12%) but only 20-40% on pages with higher baseline CER.
Truncation: on some long or error-dense pages, the model generates premature end-of-sequence tokens. A fallback mechanism (revert to original if output is >10% shorter) is recommended.
Hallucinations: the model occasionally invents corrections for proper nouns and numbers. A CER-based fallback (revert if CER worsens) catches these cases.
Corpus-specific: trained exclusively on the Viñes diaries. Performance on other HTR corpora is unknown.
Full-page only: trained on full-page text (~2500 chars). Does not generalise to shorter blocks (distribution shift).

Fallback Mechanism

For production use, always apply a fallback:

If the output is >10% shorter than the input → return original (truncation)
If ground truth is available and CER worsens → return original

This guarantees zero pages are ever worsened.

Citation

bibtex

@mastersthesis{perez2026finetuning,    title={Fine-tuning Large Language Models for HTR Post-correction:           A Case Study on the Ramon Viñes Manuscript Diaries},    author={Pérez Datsira, Albert},    school={Universitat de Lleida},    year={2026},    type={Master's Thesis}}

Acknowledgments

Carles Mateu — thesis supervisor
Esther Solé i Martí — ground truth management and HTR expertise
Màrius Bernadó — project coordination
GReia research group — Universitat de Lleida

Available on FriendliAI

Dedicated Endpoints

Run this model inference on single tenant GPU with unmatched speed and reliability at scale.

Learn more

Model Details

Model Provider

albertdatzira

Model Tree

Base

Qwen/Qwen3-8B

Adapter

this model

Input Modalities

Text

Output Modalities

Text

Supported Functionality

Dedicated Endpoints

Explore FriendliAI today

Get started Talk to an engineer

Key Result

Table with columns: Approach, CER, Pages worsened
Approach	CER	Pages worsened
HTR only (Balakirev)	10.86%	—
Llama 3.1 70B (general, no fine-tuning)	21.20%	all
This model + fallback	7.80%	0

The fine-tuned 8B model outperforms a general 70B model by 2.7× — demonstrating that domain-specific fine-tuning is essential for historical manuscript correction.

Model Details

Base model: Qwen/Qwen3-8B
Method: LoRA (Low-Rank Adaptation)
LoRA rank: 16, alpha: 32, dropout: 0.05
Trainable parameters: 43.6M (0.53% of total)
Training data: 583 page-level pairs (HTR raw → ground truth)
Training time: 56 minutes on NVIDIA DGX Spark
Epochs: 3
Language: Spanish (primary), French, Catalan

Usage

python

import torchfrom peft import PeftModelfrom transformers import AutoModelForCausalLM, AutoTokenizer # Load base model + LoRA adaptersbase_model = AutoModelForCausalLM.from_pretrained(    "Qwen/Qwen3-8B",    trust_remote_code=True,    device_map="auto",    torch_dtype=torch.bfloat16,)model = PeftModel.from_pretrained(base_model, "albertdatzira/qwen3-8b-htr-vines-lora")tokenizer = AutoTokenizer.from_pretrained("albertdatzira/qwen3-8b-htr-vines-lora") model.eval() # Correct HTR textsystem_prompt = (    "Ets un expert en post-correcció d'HTR dels diaris de Ramon Viñes. "    "Corregeix els errors de reconeixement del text manuscrit. "    "Només corregeix errors d'HTR evidents. No modifiquis l'ortografia "    "de l'autor, noms propis, ni símbols especials (¶, ¬, ⟦ ⟧).") htr_text = "para el almuerzo (á mediddia). Me hristes son todos estos aniversarios" messages = [    {"role": "system", "content": system_prompt},    {"role": "user", "content": f"Corregeix el següent text HTR:\n{htr_text}"},] text = tokenizer.apply_chat_template(    messages, tokenize=False, add_generation_prompt=True,    enable_thinking=False,)inputs = tokenizer(text, return_tensors="pt").to(model.device) with torch.no_grad():    outputs = model.generate(        **inputs, max_new_tokens=4096, do_sample=False,        pad_token_id=tokenizer.pad_token_id,    ) generated = outputs[0][inputs["input_ids"].shape[-1]:]corrected = tokenizer.decode(generated, skip_special_tokens=True)print(corrected)# -> "para el almuerzo (á mediodia). Qué tristes son todos estos aniversarios"

import torchfrom peft import PeftModelfrom transformers import AutoModelForCausalLM, AutoTokenizer # Load base model + LoRA adaptersbase_model = AutoModelForCausalLM.from_pretrained(    "Qwen/Qwen3-8B",    trust_remote_code=True,    device_map="auto",    torch_dtype=torch.bfloat16,)model = PeftModel.from_pretrained(base_model, "albertdatzira/qwen3-8b-htr-vines-lora")tokenizer = AutoTokenizer.from_pretrained("albertdatzira/qwen3-8b-htr-vines-lora") model.eval() # Correct HTR textsystem_prompt = (    "Ets un expert en post-correcció d'HTR dels diaris de Ramon Viñes. "    "Corregeix els errors de reconeixement del text manuscrit. "    "Només corregeix errors d'HTR evidents. No modifiquis l'ortografia "    "de l'autor, noms propis, ni símbols especials (¶, ¬, ⟦ ⟧).") htr_text = "para el almuerzo (á mediddia). Me hristes son todos estos aniversarios" messages = [    {"role": "system", "content": system_prompt},    {"role": "user", "content": f"Corregeix el següent text HTR:\n{htr_text}"},] text = tokenizer.apply_chat_template(    messages, tokenize=False, add_generation_prompt=True,    enable_thinking=False,)inputs = tokenizer(text, return_tensors="pt").to(model.device) with torch.no_grad():    outputs = model.generate(        **inputs, max_new_tokens=4096, do_sample=False,        pad_token_id=tokenizer.pad_token_id,    ) generated = outputs[0][inputs["input_ids"].shape[-1]:]corrected = tokenizer.decode(generated, skip_special_tokens=True)print(corrected)# -> "para el almuerzo (á mediodia). Qué tristes son todos estos aniversarios"

Training Data

The model was trained on 583 page-level pairs from the RV-training repository:

Input: raw HTR output from the Balakirev model (Kraken/eScriptorium)
Output: human-validated ground truth transcriptions
Split: by year (not random) to prevent data leakage
Filtering: pages with GT < 50 chars or CER > 80% excluded

Limitations

Best on easy pages: the model achieves 89% page improvement rate on test set pages (CER ~10-12%) but only 20-40% on pages with higher baseline CER.
Truncation: on some long or error-dense pages, the model generates premature end-of-sequence tokens. A fallback mechanism (revert to original if output is >10% shorter) is recommended.
Hallucinations: the model occasionally invents corrections for proper nouns and numbers. A CER-based fallback (revert if CER worsens) catches these cases.
Corpus-specific: trained exclusively on the Viñes diaries. Performance on other HTR corpora is unknown.
Full-page only: trained on full-page text (~2500 chars). Does not generalise to shorter blocks (distribution shift).

Fallback Mechanism

For production use, always apply a fallback:

If the output is >10% shorter than the input → return original (truncation)
If ground truth is available and CER worsens → return original

This guarantees zero pages are ever worsened.

Citation

bibtex

@mastersthesis{perez2026finetuning,    title={Fine-tuning Large Language Models for HTR Post-correction:           A Case Study on the Ramon Viñes Manuscript Diaries},    author={Pérez Datsira, Albert},    school={Universitat de Lleida},    year={2026},    type={Master's Thesis}}

@mastersthesis{perez2026finetuning,    title={Fine-tuning Large Language Models for HTR Post-correction:           A Case Study on the Ramon Viñes Manuscript Diaries},    author={Pérez Datsira, Albert},    school={Universitat de Lleida},    year={2026},    type={Master's Thesis}}

Acknowledgments

Carles Mateu — thesis supervisor
Esther Solé i Martí — ground truth management and HTR expertise
Màrius Bernadó — project coordination
GReia research group — Universitat de Lleida

qwen3-8b-htr-vines-lora

README

Key Result

Model Details

Usage

Training Data

Limitations

Fallback Mechanism

Citation

Links

Acknowledgments

Explore FriendliAI today

README

Key Result

Model Details

Usage

Training Data

Limitations

Fallback Mechanism

Citation

Links

Acknowledgments