trjxter/TraceAlchemy-Gemma-4-E4B-Finance-IT-gguf

TraceAlchemy-Gemma-4-E4B-Finance-IT GGUF

TraceAlchemy-Gemma-4-E4B-Finance-IT is a finance-focused instruction-tuned Gemma 4 E4B model trained to improve careful financial reasoning, financial table understanding, unit and scale handling, sign and direction checks, and final-answer consistency.

This repository contains the merged GGUF versions of the fine-tuned model for use with llama.cpp, LM Studio, Ollama, and other GGUF-compatible runtimes.

The model was fine-tuned, merged, and converted to GGUF using Unsloth.

Model Summary

This model was trained as a finance reasoning assistant with a focus on:

• Financial statement reasoning
• Revenue, margin, growth, and ratio calculations
• SEC-style table and excerpt extraction
• Unit and scale conversion, such as thousands to millions
• Sign and direction reasoning
• Multi-step table reasoning
• Final-answer consistency checking
• General finance instruction following

The goal of this run was not to teach the model static finance facts. Instead, the goal was to improve the model’s behavior on finance reasoning workflows. The training data emphasizes showing calculations, checking units, avoiding unsupported assumptions, and producing clear final answers.

Base Model

• Base model: unsloth/gemma-4-E4B-it
• Fine-tuned model name: TraceAlchemy-Gemma-4-E4B-Finance-IT
• Architecture class: Gemma 4 E4B instruction model
• Training method: LoRA supervised fine-tuning
• Final format in this repository: GGUF

Gemma 4 E4B is treated as an effective E4B-class model. During training, the loaded parameter count was approximately 8.1B parameters including embeddings.

Training Configuration

Setting	Value
Base model	unsloth/gemma-4-E4B-it
Training method	LoRA SFT
Base loading during training	8-bit
Max sequence length	16,384
LoRA rank	64
LoRA alpha	64
Learning rate	2e-5
Epochs	1
Per-device batch size	2
Gradient accumulation steps	8
Effective batch size	16
Optimizer	adamw_8bit
Training examples	10,000
Evaluation examples	630
Total training steps	625
Training runtime	~2.93 hours on A100-class GPU

Training Completion

The run completed successfully with the following final training output:

global_step: 625
epoch: 1.0
train_runtime: 10558.7348 seconds
train_samples_per_second: 0.947
train_steps_per_second: 0.059

The final run completed all 625 planned training steps.

Validation and Evaluation During Training

The model was evaluated every 50 training steps on a held-out evaluation set of 630 examples.

Validation loss is reported as language-modeling loss on the held-out evaluation split. It is not the same thing as benchmark accuracy, but it is useful for checking whether the model is improving on examples it is not directly training on.

The validation trend was strong during the run:

Step	Validation Loss
50	0.765309
100	0.409274
150	0.318539
200	0.286068
250	0.271857
300	0.263259
350	0.257434

The validation loss dropped from 0.765309 at step 50 to 0.257434 by step 350, showing steady improvement on held-out finance examples during training.

This suggests the model was not only fitting the training examples, but also improving on the validation set.

More benchmarks soon.

Dataset Mix

The training set used a 10,000-example finance-focused mixture.

The data recipe was built around a real finance-data anchor from previous experimentation, then expanded with targeted synthetic examples designed to address specific finance reasoning failure modes.

Real Finance Data

Source	Train	Eval
FinanceBench	120	30
TAT-QA	2,200	100
ConvFinQA	1,800	100
FinanceReasoning	750	100
Finance-Instruct-500k	400	50
Total real examples	5,270	380

Synthetic Finance Data

The synthetic examples targeted specific reasoning skills:

Synthetic Category	Train	Eval
Hard SEC-style extraction	1,500	50
Final-answer consistency checks	1,200	50
Sign, scale, and direction reasoning	900	50
Multi-step distractor tables	800	50
Hard unit-scale conversion	330	50
Total synthetic examples	4,730	250

Final Dataset Size

Split	Examples
Train	10,000
Eval	630

Dataset Credits

This model was trained using examples derived from or inspired by the following datasets:

• PatronusAI/financebench
• next-tat/TAT-QA
• AdaptLLM/ConvFinQA
• BUPT-Reasoning-Lab/FinanceReasoning
• Josephgflowers/Finance-Instruct-500k

Additional synthetic examples were generated for targeted finance reasoning skills, including SEC-style extraction, scale and unit conversion, sign and direction reasoning, multi-step table reasoning, and final-answer consistency checks.

What Was Tested

During training, the model was tested through validation loss on a held-out evaluation split containing both real and synthetic finance examples.

The evaluation set included:

• FinanceBench-style question answering
• TAT-QA table and text reasoning
• ConvFinQA conversational finance QA
• FinanceReasoning-style calculation and reasoning examples
• Finance-Instruct examples
• Synthetic SEC-style extraction checks
• Synthetic unit and scale conversion checks
• Synthetic sign and direction checks
• Synthetic final-answer consistency checks
• Synthetic multi-step distractor table checks

This validation setup was intended to check whether the model could generalize beyond the exact training examples while staying focused on finance reasoning behavior.

Further external benchmark testing is planned.

More benchmarks soon.

Available GGUF Files

File	Notes
gemma-4-E4B-it.Q4_K_M.gguf	Smaller balanced quant
gemma-4-E4B-it.Q5_K_M.gguf	Recommended balanced quality/size option
gemma-4-E4B-it.Q6_K.gguf	Higher quality quant
gemma-4-E4B-it.Q8_0.gguf	Largest quant, closest to higher precision
gemma-4-E4B-it.BF16-mmproj.gguf	Multimodal projector file for multimodal usage

For text-only use, you usually only need one of the main .gguf files, such as Q5_K_M, Q6_K, or Q8_0.

The BF16-mmproj.gguf file is for multimodal usage. It is not required for normal text-only inference.

Recommended GGUF Choice

Quant	Recommendation
Q5_K_M	Good first choice for quality/size balance
Q6_K	Better quality if you have enough VRAM/RAM
Q8_0	Highest quality among the listed quants, but larger
Q4_K_M	Smaller option when memory is limited

Example Usage

llama.cpp

llama-cli -hf trjxter/TraceAlchemy-Gemma-4-E4B-Finance-IT-gguf \
  -m gemma-4-E4B-it.Q5_K_M.gguf \
  --jinja

Multimodal llama.cpp

For multimodal usage, use the GGUF model with the mmproj file:

llama-mtmd-cli \
  -hf trjxter/TraceAlchemy-Gemma-4-E4B-Finance-IT-gguf \
  -m gemma-4-E4B-it.Q5_K_M.gguf \
  --mmproj gemma-4-E4B-it.BF16-mmproj.gguf \
  --jinja

Ollama Note

This repository includes a Modelfile generated by Unsloth for Ollama users.

The Modelfile is optional. If you are using llama.cpp, LM Studio, or another runtime that directly loads .gguf files, you can ignore it and simply download the GGUF file you want.

To create an Ollama model:

ollama create tracealchemy-gemma-finance -f ./Modelfile

Intended Use

This model is intended for finance reasoning and educational/research workflows, especially tasks involving:

• Financial statement reasoning
• SEC-style table and excerpt extraction
• Revenue, margin, growth, and ratio calculations
• Unit and scale conversion
• Sign and direction checks
• Final-answer consistency checking
• General finance instruction following

Limitations

This model should not be treated as a source of guaranteed financial truth.

It may still:

• Make calculation mistakes
• Misread financial tables
• Mis-handle units or scales
• Produce unsupported conclusions
• Hallucinate details not present in the prompt
• Fail on complex accounting, legal, or investment questions

For real investment, accounting, legal, or business decisions, verify outputs against original filings, audited statements, and qualified professionals.

Related Artifacts

This GGUF repository is part of the TraceAlchemy Gemma finance run.

Expected related repositories:

• LoRA adapter: trjxter/TraceAlchemy-Gemma-4-E4B-Finance-IT-lora
• Merged BF16 model: trjxter/TraceAlchemy-Gemma-4-E4B-Finance-IT-bf16
• GGUF model: trjxter/TraceAlchemy-Gemma-4-E4B-Finance-IT-gguf

Training and Conversion

This model was fine-tuned, merged, and converted using Unsloth.

The GGUF conversion flow was:

Base Gemma model + LoRA adapter
→ merged 16-bit model
→ BF16 GGUF
→ quantized GGUF files
→ Hugging Face upload

Generated GGUF files:

gemma-4-E4B-it.Q4_K_M.gguf
gemma-4-E4B-it.Q5_K_M.gguf
gemma-4-E4B-it.Q6_K.gguf
gemma-4-E4B-it.Q8_0.gguf
gemma-4-E4B-it.BF16-mmproj.gguf

Unsloth

This model was trained and converted with Unsloth.