This model is a full-parameter fine-tuned version of HuggingFaceTB/SmolLM-135M trained on chemistry SMILES strings from the Codemaster67/Causal_lm_chemistry_1M_rows dataset.

The base model's tokenizer was pre-extended with ~300 SPE (SMILES Pair Encoding) chemistry tokens plus <|start_of_smiles|> / <|end_of_smiles|> special tokens, and its embedding & LM-head layers were resized with mean-initialised vectors for the new tokens.

Training Details

Parameter	Value
Method	Full Fine-Tune (all weights updated)
Parallelism	FSDP (Fully Sharded Data Parallel)
Epochs	1
Learning Rate	5e-06
Batch Size (per device)	32
Gradient Accumulation	1
Max Sequence Length	128
Warmup Ratio	0.1
Weight Decay	0.01
Scheduler	Cosine
Precision	bf16
Augmentation	OFF
Training Samples	Full dataset
Eval Samples	Full dataset (10%)

Evaluation Results

Metric	Value
Final Eval Loss	3.204535961151123
Final Eval Perplexity	24.644061560759287
Training Loss	3.1543

Usage

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("Codemaster67/Test_run", trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained("Codemaster67/Test_run", trust_remote_code=True)

smiles_input = "<|start_of_smiles|>CC(=O)Oc1ccccc1C(=O)O<|end_of_smiles|>"
inputs = tokenizer(smiles_input, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=128)
print(tokenizer.decode(outputs[0], skip_special_tokens=False))

Intended Use

Chemistry-domain language modelling, SMILES generation and completion, and downstream molecular property prediction via fine-tuning.

Limitations

Trained primarily on SMILES strings; natural-language instruction-following ability may degrade compared to the base OLMo checkpoint.
Augmentation was disabled for this run.