DPLM-3B

Model overview

Synthyra/DPLM-3B packages the airkingbd/dplm_3b checkpoint with the FastPLMs runtime for Hugging Face Transformers. It accepts amino-acid sequences tokenized to masked or partially masked residue IDs.

The repository uses the standard Transformers loading interface with trust_remote_code=True. See Technical details for each registered class and whether its weights come from the checkpoint.

The sequence- and token-classification classes reuse the pretrained backbone, but their task heads are newly initialized. Fine-tune those heads before interpreting their logits as predictions.

Install and platform requirements

Install the direct dependencies published with this model:

python -m pip install -r \
  "https://huggingface.co/Synthyra/DPLM-3B/resolve/main/requirements.txt"

The FastPLMs implementation itself is embedded in the model repository. Transformers loads it through trust_remote_code=True.

This model requires Python 3.11-3.14, PyTorch 2.13, and Transformers 5.13.

The artifact requirements include the FlashAttention loader dependency. FlashAttention also requires compatible CUDA hardware and BF16 execution.

The Hub quick start needs network access for the first download. For an air-gapped run, build the manifest-pinned local artifact first and use the offline example.

Quick start

from transformers import AutoModel

model_id = "Synthyra/DPLM-3B"
model = AutoModel.from_pretrained(
    model_id,
    trust_remote_code=True,
    attn_implementation="sdpa",
).eval()

For offline validation, replace model_id with the manifest-built dist/hub/DPLM-3B path. Pass local_files_only=True.

Attention backends

The quick start uses sdpa.

Available backends are eager, sdpa, flex_attention, flash_attention_3. Requesting an unavailable backend raises instead of silently changing implementation.

output_attentions=True can use the documented one-call eager fallback to materialize attention tensors. The configured backend does not change.

Tokenization and forward inference

Load the tokenizer from the same artifact as the model. The attention mask shows padding explicitly:

import torch
from transformers import AutoTokenizer

model_id = "Synthyra/DPLM-3B"
tokenizer = AutoTokenizer.from_pretrained(
    model_id,
    trust_remote_code=True,
)
batch = tokenizer(
    ["MSTNPKPQRKTKRNT", "MKTIIALSYIFCLVFA"],
    padding=True,
    return_tensors="pt",
)

with torch.inference_mode():
    output = model(**batch)

print(output.last_hidden_state.shape)

Dataset embeddings

The shared embedding mixin keeps input order and biological-position masking. It accepts sequences, identified records, mappings, or a FASTA path:

pooled = model.embed_dataset(
    ["MSTNPKPQRKTKRNT", "MKTIIALSYIFCLVFA"],
    batch_size=2,
    pooling=("mean", "std"),
)
residues = model.embed_dataset(
    ["MSTNPKPQRKTKRNT"],
    full_embeddings=True,
)
print(pooled[0].tensor.shape)   # (2 * d,)
print(residues[0].tensor.shape) # (l, d)

Set output and format="safetensors" or "sqlite" for transactional, bounded-memory storage. Resume checks input order, model state, tokenizer policy, backend, dtype, and pooling configuration before it appends data.

Downstream prediction

The sequence and token prediction AutoClasses use the checkpoint backbone and create a new, untrained classifier. Sequence labels have shape (b,). Residue labels have shape (b, l) and use -100 outside biological positions.

import torch
from transformers import AutoTokenizer
from transformers import (
    AutoModelForSequenceClassification,
    AutoModelForTokenClassification,
)

model_id = "Synthyra/DPLM-3B"
sequence_model = AutoModelForSequenceClassification.from_pretrained(
    model_id, num_labels=2, trust_remote_code=True
).eval()
token_model = AutoModelForTokenClassification.from_pretrained(
    model_id, num_labels=3, trust_remote_code=True
).eval()
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
sequences = ["MSTNPKPQRKTKRNT", "MKTIIALSYIFCLVFA"]
batch = tokenizer(sequences, padding=True, return_tensors="pt")
biological = batch["attention_mask"].bool()
for special_id in tokenizer.all_special_ids:
    biological &= batch["input_ids"].ne(special_id)

sequence_labels = torch.zeros(len(sequences), dtype=torch.long)
token_labels = torch.full_like(batch["input_ids"], -100)
token_labels[biological] = 0

with torch.inference_mode():
    sequence_output = sequence_model(**batch, labels=sequence_labels)
    token_output = token_model(**batch, labels=token_labels)
print(sequence_output.logits.shape)  # (b, 2)
print(token_output.logits.shape)     # (b, l, 3)

PEFT fine-tuning

Install the training dependencies. Then attach LoRA to the loaded checkpoint:

python -m pip install "datasets>=4.8,<5" "peft>=0.19,<0.20"

from peft import LoraConfig, TaskType, get_peft_model

peft_model = get_peft_model(
    sequence_model,
    LoraConfig(
        task_type=TaskType.SEQ_CLS,
        r=8,
        lora_alpha=16,
        target_modules="all-linear",
        modules_to_save=["classifier"],
    ),
)

This checkpoint advertises a classification head. Save the separately trained classifier with the adapter. All FastPLMs checkpoints follow the Transformers PreTrainedModel contract and can use PEFT. The ESM2-specific shipped CLI is an example, not a support boundary. Record the target modules, base revision, data identity, and trainable parameter scope.

Test-time training

TTT samples masked views of one protein and updates only injected low-rank adapters. Base checkpoint weights stay frozen:

from transformers import AutoModelForMaskedLM

ttt_model = AutoModelForMaskedLM.from_pretrained(
    "Synthyra/DPLM-3B",
    trust_remote_code=True,
)
metrics = ttt_model.ttt(
    seq="MSTNPKPQRKTKRNT",
    ttt_config={"steps": 3, "batch_size": 1, "seed": 7},
)
ttt_model.save_pretrained("adapted", safe_serialization=True)
ttt_model.ttt_reset()
print(metrics)

Saved adapters retain their deterministic reset state. TTT adds latency and memory, can worsen an output, and does not show biological function.

Diffusion sequence generation

DPLM gets the requested length from biological positions in a tokenized input. It masks these positions and retains confident predictions at each iteration:

import torch
from transformers import AutoModelForMaskedLM, AutoTokenizer

model_id = "Synthyra/DPLM-3B"
tokenizer = AutoTokenizer.from_pretrained(model_id)
generator = AutoModelForMaskedLM.from_pretrained(
    model_id,
    trust_remote_code=True,
).cuda().eval()
input_ids = tokenizer("A" * 64, return_tensors="pt")["input_ids"].cuda()

with torch.inference_mode():
    generated_ids = generator.generate(input_ids, max_iter=100)

sequence = tokenizer.decode(
    generated_ids[0],
    skip_special_tokens=True,
).replace(" ", "")
print(sequence)

If you omit max_iter, DPLM uses the official 500-step schedule. A shorter schedule changes the sampling process. It is not an equivalent faster mode.

Plain AutoModel omits the optional ESM pooler because this diffusion checkpoint has no trained pooler weights. Pass add_pooling_layer=True only when you intend to initialize and train that head.

DPLM1 and DPLM2 checkpoint weights use Apache-2.0. The ByteDance LICENSE uses Apache-2.0. Its README limits the repository release to pretrained DPLM1 and DPLM2 weights. FastPLMs artifacts record weights_license_status="resolved" and redistributable=true. Complete publication requires all artifact, legal, parity, and atomic-publication checks.

Technical details

• Inputs: Amino-acid sequences tokenized to masked or partially masked residue IDs
• Transformers classes: AutoConfig, AutoModel, AutoModelForMaskedLM, AutoModelForSequenceClassification, AutoModelForTokenClassification
• Checkpoint weights: AutoConfig = FastPLMs extension, AutoModel = pretrained, AutoModelForMaskedLM = pretrained, AutoModelForSequenceClassification = base weights + untrained task head, AutoModelForTokenClassification = base weights + untrained task head
• Attention backends: eager, sdpa, flex_attention, flash_attention_3
• Precision: default
• BF16 execution: fp32_parameters_autocast
• Generation contract: required
• Dependencies: core
• Weight publication allowed: true
• Weight license status: resolved
• Redistributable: true
• Complete weight publication required: false

Validation and provenance

FastPLMs pins the checkpoint, upstream source revisions, state transformation, and required files in models.toml. Built artifacts record exact source identities and conversion details in source-record.json.

• FastPLMs checkpoint: Synthyra/DPLM-3B
• Runtime revision: recorded separately in the built artifact and published commit
• Runtime source identities: recorded in source-record.json
• Official checkpoint: airkingbd/dplm_3b
• Artifact source: fast
• State transform: dplm_to_fastplms_v1
• Pinned upstreams: dplm
• Release tiers: check, compliance, feature, artifact, benchmark
• Unresolved required file identities: 0

Release validation includes the compliance tier. Its evidence identifies the checkpoint, backend, dtype, hardware, inputs, and reference revision.

Declared tiers compare configuration, tokenizer behavior, state, and representative inference with the pinned reference. A nonzero unresolved count blocks release. Metadata alone does not show that a build passed, that a backend is faster, or that an output is biologically valid.

License

Checkpoint terms: Apache-2.0. The Hub model-card identifier is apache-2.0. The local artifact contains applicable source licenses, notices, attribution, and conversion records. Review them before use.