GitHub - moritztng/tt-bio: Run protein structure prediction, binder design, and embeddings on Tenstorrent hardware — Boltz-2, ESMFold2, Protenix-v2, OpenFold3, OpenDDE, BoltzGen, RFdiffusion3, ESMC, SaProt.

████████╗████████╗        ██████╗  ██╗  ██████╗
╚══██╔══╝╚══██╔══╝        ██╔══██╗ ██║ ██╔═══██╗
   ██║      ██║    █████╗ ██████╔╝ ██║ ██║   ██║
   ██║      ██║    ╚════╝ ██╔══██╗ ██║ ██║   ██║
   ██║      ██║           ██████╔╝ ██║ ╚██████╔╝
   ╚═╝      ╚═╝           ╚═════╝  ╚═╝  ╚═════╝

Important

TT-Boltz is now TT-Bio

TT-Bio runs Boltz-2, ESMFold2, Protenix-v1 and Protenix-v2, OpenFold3, OpenBind-0, OpenDDE, and RoseTTAFold3 structure prediction, BoltzGen, RFdiffusion3 and PXDesign binder/protein design, and ESMC protein embeddings, SaProt structure-aware protein embeddings, and Nesso-1 structure-free binding affinity on Tenstorrent Blackhole and Wormhole, supporting single-card and multi-card configurations (e.g. QuietBox with 4 cards or Galaxy server with 32 cards). Across several machines, each runs its own controller and a scheduler of your choice spreads the work; see Running tt-bio on many machines.

Benchmarks: tt-bio.com has throughput and cost for every model against NVIDIA DGX H200, B200 and A100. The full benchmark page has the measured seconds behind every figure, the fixtures they were run on, the run conditions and the cost model.

Accuracy

Every model TT-Bio serves is validated against its official reference implementation on the same input and reproduces it within that reference's own run-to-run noise. See docs/implementation-parity.md for the methodology, per-target results, and reproduction commands.

Performance and cost

Predictions and designs per hour per server, and throughput per dollar of purchase price and of total cost of ownership, measured against NVIDIA DGX systems. See the benchmark page for the numbers, the fixtures they were measured on, and the cost model behind them.

Installation

Create a Python virtual environment with Python 3.10 or 3.12, install with the Tenstorrent extra, then install the matching Tenstorrent system dependencies.

We test on Ubuntu 24.04 (glibc 2.39, Python 3.12). The ttnn wheel is tagged manylinux_2_34 and imports GLIBC_2.34 symbols, so glibc 2.34 or newer is required: RHEL 8 and its rebuilds ship glibc 2.28 and cannot install it. See Troubleshooting for the CPU frequency driver check, which matters more than anything else about the host.

python3.10 -m venv env
source env/bin/activate
pip install 'tt-bio[tenstorrent]'
tt-bio install-deps

tt-bio install-deps installs the Tenstorrent system dependencies that match this release. It may ask for your sudo password.

On a host without a Tenstorrent card, plain pip install tt-bio is enough: the Boltz-2 CPU/GPU path (--accelerator cpu / --accelerator gpu) and the CLI work without the Tenstorrent SDK. The other models run on Tenstorrent only.

From GitHub / source

Pin to a tagged release, track nightly main (may be untested), or work from an editable clone:

pip install "tt-bio[tenstorrent] @ git+https://github.com/moritztng/tt-bio.git@v0.9.0"   # pinned release, see Releases for the latest
pip install "tt-bio[tenstorrent] @ git+https://github.com/moritztng/tt-bio.git@main"     # nightly
# or
git clone https://github.com/moritztng/tt-bio.git
cd tt-bio
pip install -e '.[tenstorrent]'
tt-bio install-deps

Drop the [tenstorrent] extra on a host without a Tenstorrent card.

Optional: Build TT-Metal / TT-NN from Source

If you need to build from source, follow the Tenstorrent Installation Guide.

Verify Installation

tt-bio --version   # or -V; prints the installed version
tt-bio --help
tt-bio predict --help
tt-bio msa --help

Troubleshooting

Check the CPU frequency driver first. On an AMD host:

cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver   # want: amd-pstate-epp

If it reads acpi-cpufreq, every ttnn op pays a fixed extra host cost, measured at roughly 5 to 12 microseconds, and throughput drops across every model. The symptom is distinctive: per-op device time is unchanged and the loss tracks a model's op count rather than its size, so a fold that issues 25k ops loses far more than an embedding that issues 1k. One user measured Boltz-2 at 1.29 structures/s under a 5.4 kernel with acpi-cpufreq and 2.14 structures/s on the same box, same binaries, after booting a kernel that gives amd-pstate-epp.

amd-pstate-epp needs Linux 6.3 or newer. A kernel configured with CONFIG_X86_AMD_PSTATE_DEFAULT_MODE=3 selects it with no boot flag; Ubuntu 24.04 does this. On an older kernel that has the driver but does not default to it, add amd_pstate=active to the kernel command line. Kernels before 6.3 (RHEL 8, Oracle UEK 6, anything 4.18 or 5.4) have no amd-pstate at all.

Single-host prediction needs no MPI setup of yours. tt-metal ships the OpenMPI it wants, and a system OpenMPI on the environment breaks that build instead of replacing it: with OMPI_MCA_* or OPAL_PREFIX set, or a foreign libmpi on LD_LIBRARY_PATH, MPI_Init aborts before the fold starts. tt-bio warns when it sees one and changes nothing for you. To clear it for the current shell:

unset OMPI_MCA_pml OMPI_MCA_plm OPAL_PREFIX LD_LIBRARY_PATH

Basic Usage

Structure Prediction

tt-bio predict examples/prot.yaml --model boltz2 --override

Every command names its model with --model:

boltz2: folds complexes of proteins, DNA, RNA, and ligands and predicts binding affinity. MSA-dependent (uses an MSA by default).
esmfold2 / esmfold2-fast: fold complexes of proteins, DNA, RNA, and ligands (CCD code or SMILES) on-device, no MSA required (esmfold2-fast is the lighter, faster checkpoint). Modified residues are folded as the modified chemistry, and covalent bond constraints and cyclic peptides are supported. Templates and pocket constraints are refused.
protenix-v1 / protenix-v2: fold complexes of proteins, RNA, DNA, and ligands (an AlphaFold3-family model, the Protenix reproduction); MSA-dependent for proteins (uses an MSA by default), and also emit a PAE/PDE matrix with --write_pae. protenix-v1 is upstream's own v0.5.0 base checkpoint: half the pair width and 4 trunk recycles against protenix-v2's 10, so it is the cheaper of the two. Covalent bond constraints and modified residues are supported on both; cyclic peptides and templates, as a structure file or a precomputed alignment (see Templates), on protenix-v2 only. protenix-v1's checkpoint ships no template blocks, and it does not close a cyclic peptide's ring.
openfold3: folds proteins, RNA and DNA (an AlphaFold3-family model, the OpenFold3 reproduction); MSA-dependent (uses an MSA by default), with optional templates, as a structure file or a precomputed alignment (see Templates). Polymer chains only, so ligands are refused. Cyclic peptides, modified residues and a covalent bond to a modified residue are supported; a bond between two standard residues (a disulfide) is refused. Weights come from the OpenFold consortium; point OF3_CKPT at them.
openbind: OpenBind-0, the same OpenFold3 stack on upstream's v0.5.0 checkpoint, tuned for protein-ligand co-folding. Takes ligands by SMILES or CCD code alongside protein, RNA and DNA chains; MSA-dependent (uses an MSA by default), with optional templates, as a structure file or a precomputed alignment (see Templates). Cyclic peptides, modified residues and covalent bonds to a ligand or modified residue are supported; a bond between two standard residues (a disulfide) is refused. Weights are a separate file from openfold3 and are not downloaded; point TT_BIO_OPENBIND at them (see docs/weights.md).
saprot: structure-aware protein embeddings, an ESM-2 encoder over a fused amino-acid + Foldseek-3Di vocabulary (446 tokens). Needs a structure for the 3Di structural tokens (--structure); runs sequence-only without it. Use for variant-effect / mutation-fitness scoring and function prediction.
nesso1 (tt-bio affinity): protein-ligand binding affinity without a structure. Predicts a soft distogram and reads the affinity off that, so it is much cheaper than folding and it returns no coordinates. Proteins and ligands only.
opendde / opendde-abag: antibody-antigen co-folding built on the Protenix-v2 stack plus a structural-token expander; opendde-abag selects the antibody-antigen checkpoint. Protein, RNA, DNA and ligand chains, with covalent bond constraints, cyclic peptides, modified residues and templates in either form (see Templates). Proteins are MSA-dependent (uses an MSA by default, like Protenix-v2).
af2ig: AlphaFold2 initial-guess, the filter binder-design pipelines use to tell a real design from a plausible one. It takes a design you already have -- a structure carrying the target chain and the binder backbone, plus the binder's sequence -- re-predicts the complex starting from those coordinates, and reports pLDDT, pTM, ipTM, pAE and interface pAE. Single-sequence, no diffusion, no seed: the same input gives the same answer. Input format and an example are in examples/af2_designed_complex.yaml; the file takes target: and binder: and nothing else, and any other key is refused rather than ignored. Weights are DeepMind's AlphaFold2 monomer pTM parameters; tt-bio weights --download af2ig fetches them (4 GB, one file kept).
rf3: folds complexes of proteins, RNA, DNA, and ligands (an AlphaFold3-family model, RoseTTAFold3 from the Institute for Protein Design); MSA-dependent for proteins (uses an MSA by default). Writes AlphaFold3-style <name>_summary_confidences.json (pTM, ipTM, chain-pair PAE/PDE, ranking score) next to each structure. Modified residues and covalent bonds to a ligand or modified residue are supported; cyclic chains, a bond between two standard residues (a disulfide) and pocket constraints are refused. Templates work in either form (see Templates). Weights download from the IPD on first use.

tt-bio predict examples/prot.fasta --model esmfold2-fast --fast
tt-bio predict examples/fkg_ligand.yaml --model esmfold2   # protein + ligand co-fold, no MSA
tt-bio predict examples/prot.yaml --model protenix-v2   # MSA on by default; NA/ligand chains are single-sequence
tt-bio predict examples/prot.yaml --model protenix-v1   # upstream's v0.5.0 base checkpoint, 4 recycles
tt-bio predict examples/prot.fasta --model openfold3    # MSA on by default; set OF3_CKPT to the weights file
tt-bio predict examples/affinity.yaml --model openbind  # protein + ligand co-fold; set TT_BIO_OPENBIND
tt-bio predict examples/9dsg_abag.yaml --model opendde-abag   # antibody-antigen co-fold, MSA on by default
tt-bio predict examples/prot.yaml --model rf3            # MSA on by default; weights fetch from the IPD
tt-bio predict examples/prot.yaml --model rf3 \
    --partial_t 150 --partial_structure start.cif       # refine start.cif instead of folding from scratch
tt-bio predict targets.yaml --model rf3 --early_stop_plddt 0.5   # skip the rollout on hopeless targets
tt-bio predict examples/af2_designed_complex.yaml --model af2ig  # score a designed complex: ipTM + interface pAE

Feature	Boltz-2	ESMFold2	Protenix-v1	Protenix-v2	OpenFold3	OpenBind-0	OpenDDE	RF3
Input	protein/DNA/RNA/ligand complex	protein/DNA/RNA/ligand complex	protein/DNA/RNA/ligand complex	protein/DNA/RNA/ligand complex	protein/RNA/DNA (polymer-only)	protein/DNA/RNA/ligand complex	protein/DNA/RNA/ligand complex	protein/DNA/RNA/ligand complex
MSA	MSA-dependent (on by default)	single-sequence	proteins MSA-dependent (on by default), NA/ligand single-sequence	proteins MSA-dependent (on by default), NA/ligand single-sequence	proteins MSA-dependent (on by default)	proteins MSA-dependent (on by default)	proteins MSA-dependent (on by default), NA/ligand single-sequence	proteins MSA-dependent (on by default)
PAE/PDE output (`--write_pae`)	no	no	yes	yes	no	no	no	in `_summary_confidences.json`

Ligands, nucleic acids, modified residues, cyclic chains, covalent and pocket constraints, templates and affinity are one generated matrix: docs/model-capabilities.md. Anything a model cannot honour is refused by name before the fold starts, never accepted and dropped.

Every structure model folds at least 1536 residues on a single 12 GiB Wormhole chip. The limit below is the largest size that folded, under the first measured failure where one was found, walked with the settings the platform sends:

model	Wormhole limit	first measured failure
`boltz2`	1920	2048
`opendde`, `opendde-abag`	1536	1664 (it runs for hours rather than crashing)
`openfold3`	1664	1792
`openbind`	1664 (residues; a ligand adds tokens)	1792
`pxdesign`	1536 (target residues; the binder is on top)	none found; top of the ladder
`protenix-v2`	2048 (residues; a ligand adds tokens)	none found; top of the ladder
`esmfold2`	1664 (residues; a ligand adds tokens)	1792
`esmfold2-fast`	1664 (residues; a ligand adds tokens)	1792
`rf3`	1600	1664
`protenix-v1`	2048	none found; top of the ladder
`rfd3`	1536 (motif + designed)	none found; top of the ladder
`boltzgen`	14786 (atoms in the target)	none found; top of the ladder
`esmc-6b` (embed)	1968; 8192 with `--fast`	1984

For every model here that reads an alignment, the limit was measured with 16384 alignment rows, the most any of them reads, so a deeper a3m does not lower it.

Ask for more than a model's limit and tt-bio refuses before it opens a device, naming the model, the limit and any model that does take the input. nesso1 and af2ig have no measured limit and are never refused: nothing above the top of either ladder has been run, so there is no failing size to refuse on. AF2-IG's top rung is 1024 tokens (944 target residues plus an 80-residue binder) in 488 s on one Wormhole chip, and 1536 tokens (1008 plus a 528-residue binder) in 665 s on a Blackhole p150a. What sets each wall is in docs/large-targets.md.

boltzgen is the one model sized on atoms rather than residues, because its wall follows the target's atom count and atoms per residue vary with what the target is made of: the 14786-atom rung is 1831 residues of a deposited protein, and a lighter target reaches more residues in the same number of atoms. The designed binder is outside the number, which was 80 residues at every rung. tt-bio counts the atoms out of the structure file your file: entity points at, so the refusal still lands before a device opens.

Blackhole has more memory per chip, so it is a separate table and the Wormhole numbers do not carry over. Six models have been walked to a failure on a p150a:

model	Blackhole limit	first measured failure
`saprot-35m` (embed)	126976 (longest sequence)	131072
`esmc-300m`, `esmc-600m`, `saprot-650m`, `saprot-1.3b` (embed)	114688 (longest sequence)	126976
`pxdesign`	2500 (target residues)	3000

The embedding wall is one allocation: the model asks for the whole sequence-by-sequence attention matrix as a single buffer that grows with the square of the sequence length, 34.4 GB at 131072 residues against a 31.9 GiB card. saprot-35m gets one rung further than the rest because its weights are the smallest and leave more room for it. Everything else is unmeasured on Blackhole and refused nothing, including boltzgen, which designs against a 16948-atom target there.

The Blackhole ladders have been walked separately. This is the largest rung each structure model is on record folding, from docs/size_ladder_baseline.d/:

model	p150a	p300c
`boltz2`	1536	1024
`esmfold2`	1536	1024
`nesso1`	1536	1024
`openbind`	1536	1024
`opendde`	1024	1024
`openfold3`	1536	1024
`protenix-v1`	1536	1024
`protenix-v2`	1536	1024
`rf3`	1536	1088

Only opendde and opendde-abag are enforced on Blackhole, at 1024; every other model here accepts an oversized request rather than refusing it, so the table is what you have and not a guard. A blank rung above a model's top is unwalked, not a measured failure -- 1536 is the top of the ladder itself, so the models sitting there have no failure on record at all.

The limits were measured with an MSA, which is the default for the models that take one, and at the deepest alignment the MSA pipeline actually produces. Folding single-sequence is roomier, so if you know your run is lighter than the ladder that set the limit, TT_BIO_SIZE_LIMIT=0 turns the refusal into a warning and runs it anyway.

A ligand counts against these limits. Its heavy atoms are tokens the model pays for exactly like residues, and on esmfold2, esmfold2-fast, openbind and protenix-v2 the wall is on tokens, so a cocrystal is checked on residues plus ligand atoms rather than on the residue count alone. esmfold2 folds 1664 residues, which leaves no room at all: 1664 residues plus any ligand is refused, and 1640 residues with a 24-atom ligand is admitted. openbind is the same: its 1664 was walked apo, so a ligand counts against it atom for atom. Either way the refusal names the token count and the wall, and it arrives before a device is opened instead of as an out-of-memory error part way through the fold.

esmfold2-fast is the same architecture at half the trunk depth. It has its own row because it was walked on its own; today both fold 1664 residues and fail at 1792.

The pair track switches to row-blocked execution at a size threshold smaller targets never reach, so their speed and numerics are untouched. See docs/large-targets.md. Perf levers are gated at several sequence lengths, not just one; the release gate re-checks the ladder against a recorded baseline. See docs/size-generality.md.

All structure models support the sampling, output-format, and scheduling options. MSA, affinity, constraint, and auxiliary-output options apply only where listed below. Each model downloads its weights automatically on first use, except OpenFold3: fetch the consortium checkpoint yourself and point OF3_CKPT at it, or put it at ~/.boltz/of3-p2-155k.pt. tt-bio weights lists every artifact with its status, size and path; --download prefetches, --prune reclaims disk. Set TT_BIO_CACHE to move all of it (both ~/.boltz and the Hugging Face cache, about 65 GiB) somewhere with room. See docs/weights.md.

Boltz-2, Protenix-v1, Protenix-v2, OpenFold3, OpenDDE, and RF3 are MSA-dependent and use an MSA by default, a local ColabFold DB (~/.boltz/msa_db) if one is set up (see Offline MSA), otherwise the online ColabFold server. Sending sequences to the online server (api.colabfold.com) leaves your machine; a one-line notice is printed when that fallback is used. Pass --msa_db_path for a private offline database, or --single_sequence to deliberately fold without an MSA (lower accuracy; for batch-screening orphan sequences). A complex with two or more different protein sequences also gets a species-paired MSA, searched once per complex, the way each model's upstream pairs; a homodimer is not paired. Two models fold unpaired because their upstream does: RF3 pairs by taxonomy IDs that ColabFold alignments do not carry, and OpenFold3's preview2 checkpoint runs on an upstream release that drops the paired rows (OpenBind pairs). ESMFold2 needs no MSA and uses one when a source is given.

--fast makes some operations use a lower-precision numeric format that runs faster. Accuracy is typically very close.

OpenDDE-abag matches the upstream checkpoint on the standard 1AHW antibody-antigen target. Both implementations perform poorly on 9DSG.

predict accepts either a single YAML/FASTA file or a directory containing many input files. An input the model refuses is reported, recorded as failed in results.json and skipped, and the rest of the directory folds; the exit status is 2 when some inputs failed and 1 when all did. Every chain comes back under the id you gave it.

A live display shows the progress of each target. Prediction uses up to one card per pending target, labelled in the display (quietbox:tt0, quietbox:tt1, ...). Models load once per active card and stay resident:

tt-bio predict proteins/ --model boltz2 --out_dir results --fast

Pass --devices 0,1,2,3 to pick or limit the available cards. A single target remains a single-card fold; additional cards increase throughput only when multiple targets are queued.

Once enough folds run at once that each worker is down to a couple of CPU threads, tt-bio also has their idle thread pools sleep between device syncs instead of spinning on them. The output is identical either way, and a single fold is unaffected. See Tuning flags for the measurement, or set OMP_WAIT_POLICY yourself to take the decision back.

To spread work across several machines, see Running tt-bio on many machines.

Protein Embeddings (ESMC)

Turn protein sequences into ESMC language-model embeddings on-device (no folding, no MSA). DATA is a FASTA file, a directory of them, a YAML {id: sequence} mapping, or a bare sequence string:

tt-bio embed proteins.fasta --model esmc-600m --out_dir embeddings
tt-bio embed "MQIFVKTLTGKTITLEV..." --model esmc-600m   # one-off sequence

--model selects the ESMC variant (esmc-300m, esmc-600m, esmc-6b). For each sequence you get its per-residue embeddings ([length, d_model] float32, one row per amino acid, row order == input order) and a pooled whole-sequence vector ([d_model] float32, --pool mean/max/cls). --out_dir (default ./embeddings) gets:

<id>.npz per sequence: per_residue, pooled (+ logits with --logits); --format npz, default
embeddings.parquet: pooled vectors, one row per sequence; --format parquet
manifest.json: model/pool/shapes/dtype and which file holds each sequence

Add --logits for the per-residue amino-acid predictions (300M/600M only), and --fast for the lower-precision weight path. Weights download automatically on first use.

Sequences batch automatically on 300M/600M (--batch_size, default 8): a padded, length-bucketed device forward per batch, masked so results are identical to running each sequence alone. Single-sequence calls (--batch_size 1, e.g. serving one sequence at a time) replay through a captured device trace once a length bucket repeats, up to ~1.5x faster per call on QuietBox-class hosts, bit-identical, no flags needed.

To embed a large batch faster, shard it across several cards with --devices 0,1,2,3: one worker per card, results reassembled in input order and identical to a single-card run:

tt-bio embed proteins.fasta --model esmc-600m --devices 0,1,2,3

Measured, not assumed: fanout only pays off when there's enough work per shard to amortize each worker's model-load and device-init cost. On small batches it can be flat or worse than a single card. esmc-6b scales to 4 cards on suitably large batches. Reach for --devices on large batches, not small ad-hoc jobs; use --controller (below) for repeated/production embedding.

For repeated/production embedding, submit to a persistent pool instead: a worker loads its model once and keeps it resident across every call, so the reload cost above is paid once per worker, not once per invocation:

tt-bio controller --port 8765            # starts + keeps a worker per local card
tt-bio embed proteins.fasta --model esmc-6b --controller http://127.0.0.1:8765

The same capability is available from Python:

from tt_bio import esmc

emb = esmc.embed("MQIFVKTLTGKTITLEV...", model="esmc-600m")[0]
emb.per_residue   # [L, d_model] float32
emb.pooled        # [d_model] float32

# Shard a large set across cards (data-parallel, order preserved):
embs = esmc.embed(sequences, model="esmc-600m", devices=[0, 1, 2, 3])

Structure-Aware Protein Embeddings (SaProt)

SaProt is a structure-aware protein language model, an ESM-2 encoder over a fused amino-acid + Foldseek 3Di vocabulary (446 tokens). Where ESMC is sequence-only, SaProt also encodes local structure, so its embeddings and MLM logits reflect both sequence and shape. Use it for variant-effect / mutation-fitness scoring and function prediction when you have a structure (predicted or experimental: fold it with tt-bio predict first, then score it with SaProt).

tt-bio saprot proteins.fasta --model saprot-650m --structure structs/ --out_dir embeddings
tt-bio saprot proteins.fasta --model saprot-650m                # sequence-only (3Di = '#')
tt-bio saprot proteins.fasta --model saprot-650m --devices 0,1    # data-parallel across 2 cards

--structure is a PDB/cif file (single sequence) or a directory of <id>.pdb/<id>.cif files, one per FASTA id. The 3Di structural tokens are computed on host with Foldseek (conda install -c bioconda foldseek; --foldseek PATH or FOLDSEEK_BIN if it is not on PATH); it runs off-device. Residues the structure does not resolve get the # unknown-structure token, and a structure that is not of the sequence you passed is refused rather than lined up by length. Omit --structure for sequence-only mode (lower accuracy for 35M/650M; the 1.3B works sequence-only).

For each sequence you get per-residue structure-aware embeddings ([length, d_model] float32) and a pooled vector, plus per-residue MLM logits ([length, 446] with --logits) over the fused vocabulary, the log-likelihoods used for zero-shot mutation scoring. Output layout matches tt-bio embed (<id>.npz / embeddings.parquet / manifest.json).

--model selects the variant (saprot-35m, saprot-650m, saprot-1.3b). --devices 0,1,2,3 shards the input across cards data-parallel (one pinned subprocess each, results reassembled in input order), bit-exact vs single-card with --batch_size 1. Parity vs the reference HuggingFace checkpoint, the multi-card bit-exactness check, and warm throughput are in docs/saprot-parity.md.

Python:

from tt_bio import saprot

emb = saprot.embed(("MQIFVKTLTGKTITLEV...", "dweweaepvrdidi..."), model="saprot-650m")[0]
emb.per_residue   # [L, d_model] float32, structure-aware
emb.logits        # [L, 446] float32 (with return_logits=True)

Weights

Weights download on first use, so nothing here is required. tt-bio weights is for when you want to see or move them:

tt-bio preflight protenix-v1         # can this machine run it right now?
tt-bio weights                       # every artifact: status, size, resolved path
tt-bio weights --download            # prefetch everything (e.g. before going offline)
tt-bio weights --download boltz2     # or just one model's set
tt-bio weights --prune               # reclaim superseded revisions and leftovers

tt-bio preflight answers before you submit a job, and exits non-zero when something is missing, so it works in a script. When weights are missing it also measures whether the hosts they come from can be reached from this machine.

A full set is about 65 GiB. It lands in ~/.boltz and the Hugging Face cache; set TT_BIO_CACHE to put both somewhere with more room. Each artifact also takes its own override, so TT_BIO_BOLTZ2_CONF=/mnt/weights/boltz2_conf.ckpt loads that file instead of downloading. Rows show as corrupt if a download was interrupted, and are re-fetched rather than loaded. No download waits forever: a source that sends nothing is dropped for the next one, and the error names every host tried. See docs/weights.md.

Offline MSA (Optional)

Use this if you have enough disk and RAM and want local MSA. This avoids external MSA server calls and is faster for repeated runs.

tt-bio msa
tt-bio predict examples/prot.yaml --model boltz2 --override

tt-bio msa downloads UniRef30 to ~/.boltz/msa_db (~100GB download, ~500GB on disk after indexing). predict auto-detects this path.

To add EnvDB and use it in prediction: EnvDB can improve MSA coverage when UniRef30 hits are weak, at higher disk/RAM cost.

tt-bio msa --db all
tt-bio predict examples/prot.yaml --model boltz2 --use_envdb --override

Key Options:

--override: Re-run from scratch, ignoring cached files
--use_msa_server: Generate MSA via ColabFold API
--msa_db_path: Use a local database at a custom path (e.g. --msa_db_path /data/colabfold_db)
--use_envdb: Include EnvDB in offline MSA (tt-bio msa --db all)
--accelerator=tenstorrent: Use Tenstorrent hardware (default, or use cpu/gpu)
--fast: Makes some operations use a lower-precision numeric format that runs faster; accuracy is typically very close
--debug: Show all raw output from the hardware and libraries instead of the progress display
--debug --log: Same as --debug, but also print what each device is currently working on

Shared MSA Server (Optional)

Host the database on one machine and let others fetch MSAs from it over HTTP, so each prediction machine need not keep its own ~500GB copy.

# On the machine with the database:
tt-bio msa-server --listen 0.0.0.0:8765

# On any other machine (no local database needed):
tt-bio predict examples/prot.yaml --model protenix-v2 --msa_endpoint http://HOST:8765

The server runs the same offline colabfold_search and serves unpaired {hash}.a3m, with a shared cache and a search-concurrency cap (--max_concurrent). Add --token to require Authorization: Bearer <token>. --msa_endpoint applies to --model esmfold2, protenix-v1, protenix-v2, openfold3, opendde, and rf3.

Binding Affinity Prediction (Boltz-2)

Predict binding affinity for protein-ligand complexes:

tt-bio predict examples/affinity.yaml --model boltz2 --use_msa_server --override --affinity_mw_correction

The --affinity_mw_correction flag applies molecular weight correction for more accurate predictions.

An affinity run folds the complex and then runs a second model that has its own 64-block trunk, so it costs more than a structure-only fold. All of it runs on the card. FKBP12+SB3 at the default affinity protocol (200 sampling steps, 5 affinity samples, single sequence) takes about 206 s per ligand on one Blackhole p150a, measured as a whole tt-bio predict invocation with model load included. --sampling_steps_affinity and --diffusion_samples_affinity are the two flags that move that wall most.

The affinity trunk runs in fp32 because the predicted log10(IC50) is sensitive to activation precision, and that is not configurable. Earlier releases ran it in fp32 on the host CPU instead, which is why affinity used to take minutes per ligand and looked CPU-bound.

Binding Affinity Without a Structure (Nesso-1)

tt-bio affinity predicts protein-ligand affinity without folding anything. Nesso-1 has no structure module, so it returns a number, not coordinates:

tt-bio affinity examples/affinity.yaml                  # one complex
tt-bio affinity ligands/ --out_dir screen               # a directory is a screen

A directory keeps the model resident across inputs, so a ligand series against one target pays the weight load and the kernel compile once. Output is one <id>_affinity.json per input plus an affinity.csv for the whole run: the affinity value (mean of a two-member ensemble, and each member), a binary binder probability, and six distogram entropies.

On DAVIS it reaches 0.662 mean within-target Pearson against measured Kd (0.175 for a molecular-weight-only control), matching the 0.636 the upstream implementation gets on an H200.

It is far cheaper than folding for the same question. One 512 aa prediction takes 8.3 s of model time on one Blackhole card, 33 s for the whole command including featurisation; Boltz-2 affinity takes 386 s for the same command on the same input, which is what tt-bio shipped for this before. Against a GPU it is 7.9x off an H200 at that size, so choose it for what the answer costs on this hardware rather than expecting it to beat a GPU.

Use it to rank a series; use predict --model boltz2 when you need the pose. Proteins and ligands only, one ligand scored per input. One Wormhole chip scores a 3072-residue target with any ligand up to cobalamin's size in about 15 minutes. 3584 residues also completes but takes about two hours, because the chip's memory spills to the host; 4096 is refused. The trunk runs bf16 by default: it is about 6x faster than fp32 and no less accurate from 276 tokens up, and fp32 runs out of DRAM around 1000 tokens. On inputs under ~150 tokens fp32 is the more faithful arm, and --trunk fp32 switches back. See docs/nesso1.md for the input schema, the four upstream limits, and what to watch when comparing numbers against another implementation.

Input Format

ESMFold2, Protenix-v1 and Protenix-v2 accept proteins, DNA, RNA, ligands and covalent bond constraints. OpenFold3 accepts proteins, DNA and RNA plus per-chain templates, and refuses ligands. OpenDDE accepts proteins and ligands with bond constraints. Boltz-2 additionally supports affinity, pocket/contact constraints, potentials, and user-supplied templates. Which model takes cyclic chains and which kind of bond is in docs/model-capabilities.md.

Create a YAML file describing your complex:

version: 1
sequences:
  - protein:
      id: A
      sequence: MVTPEGNVSLVDESLLVGVTDEDRAVRSAHQFYERLIGLWAPAVMEAAHELGVFAALAEAPADSGELARRLDCDARAMRVLLDALYAYDVIDRIHDTNGFRYLLSAEARECLLPGTLFSLVGKFMHDINVAWPAWRNLAEVVRHGARDTSGAESPNGIAQEDYESLVGGINFWAPPIVTTLSRKLRASGRSGDATASVLDVGCGTGLYSQLLLREFPRWTATGLDVERIATLANAQALRLGVEERFATRAGDFWRGGWGTGYDLVLFANIFHLQTPASAVRLMRHAAACLAPDGLVAVVDQIVDADREPKTPQDRFALLFAASMTNTGGGDAYTFQEYEEWFTAAGLQRIETLDTPMHRILLARRATEPSAVPEGQASENLYFQ
  - ligand:
      id: B
      smiles: 'N[C@@H](Cc1ccc(O)cc1)C(=O)O'
properties:
  - affinity:
      binder: B

Entity Types:

Polymers (protein, dna, rna): provide sequence
Ligands (ligand): provide smiles or ccd code

Multiple Identical Chains:

- protein:
    id: [A, B]  # Two identical chains
    sequence: ...

Understanding Results

Output Structure

<model>_results_prot/   # e.g. protenix_results_prot, boltz2_results_prot
├── structures/
│   ├── prot.cif                      # Best-ranked predicted structure
│   └── prot_model_1.cif              # Additional samples (if diffusion_samples > 1)
├── results.json                      # One entry per target with confidence/affinity metrics
├── power_profile.csv                 # (optional, --report-energy)
├── power_profile.png                 # (optional, --report-energy)
├── prot_pae.npz                      # (optional, --write_pae)
├── prot_plddt.npz                    # (optional, --write_pae, Boltz-2)
├── prot_pde.npz                      # (optional, --write_pde)
└── prot_embeddings.npz               # (optional, --write_embeddings)

MSA results are cached in <out_dir>/msa/ (default ./msa/), keyed by sequence hash. The same protein sequence is never searched twice, even across different input files or runs. The MSA search uses all available CPU threads and keeps the database index memory-mapped for maximum speed.

Confidence Scores

Each target entry in results.json contains confidence metrics. The fields below are Boltz-2's; Protenix-v2 and OpenFold3 report the same confidence_score / ptm / iptm / plddt (and all_runs when --diffusion_samples > 1, ranked best-first), while an ESMFold2 entry instead carries plddt (mean, 0-1), ptm when available, and n_residues / n_chains. Every model reports its complex mean pLDDT under plddt. Boltz-2, Protenix-v2 and OpenDDE also report the two per-chain fields below on a multi-chain target, in all_runs as well as for the best sample.

{
    "id": "prot",
    "status": "ok",
    "confidence_score": 0.84,
    "ptm": 0.84,
    "iptm": 0.82,
    "complex_plddt": 0.84,
    "plddt": 0.84,
    "chains_ptm": {
        "0": 0.85,
        "1": 0.83
    },
    "pair_chains_iptm": {
        "0": {"0": 0.85, "1": 0.72},
        "1": {"0": 0.82, "1": 0.83}
    }
}

confidence_score: Overall confidence (0-1, higher is better). Models are ranked by it and the best sample is written as <name>.cif. Boltz-2 and BoltzGen use their own published rule, 0.8 × complex_plddt + 0.2 × iptm, or 0.2 × ptm on a single chain where there is no interface and iptm is 0; it is not a pLDDT and can sit either side of one, on CDK2 0.008 above complex_plddt at 298 aa and 0.050 below it at 512 aa. The AlphaFold3-family models (OpenFold3, OpenBind-0, RF3, Protenix-v1/v2, OpenDDE) share one rule: AlphaFold3's 0.8 × iptm + 0.2 × ptm + 0.5 × disorder − 100 × clash, with plddt taking ipTM's weight on a single chain, where ipTM is zero by construction. OpenFold3 and OpenBind-0 are the only two that compute the disorder and clash terms; the others pass 0 for them. RF3 reports the same number as ranking_score. So the family's scores are comparable to each other but not to Boltz-2's
ptm: Predicted TM-score for complex (0-1)
iptm: Interface TM-score (0-1)
complex_plddt, plddt: Mean confidence (0-1), the same value under both names. It is the mean of the B-factor column of the structure file the same fold wrote, so averaging that column reproduces it. Boltz-2 writes one pLDDT per residue, so average over one atom per residue (CA); Protenix-v2, OpenFold3, OpenBind-0 and OpenDDE write one per atom, so average over all of them
chains_ptm: Per-chain TM-scores (0-1)
pair_chains_iptm: Per-chain-pair interface TM-scores (0-1), with each chain's own chains_ptm on the diagonal. Read pair_chains_iptm[binder][target] to score one named interface of a complex; the global iptm is the whole-interface number and on a two-chain target the two agree. Like every other confidence value, these are comparable between targets of the same model, not between models

Interface Scores

With --write_pae, a multi-chain Boltz-2 entry also carries interface_scores: ipSAE (both directions, their max and their min), ipTM, interface pAE, pDockQ, pDockQ2 and LIS for every chain pair. The definitions match the script Adaptyv scored its Nipah binder competition with, at the same 15 A cutoffs. interface_score_distribution gives the same scores for every diffusion sample, with mean, standard deviation and range, so --diffusion_samples 5 shows how stable a design's score is as well as its value. To score a fold you already have, from tt-bio or from upstream Boltz, without a device:

tt-bio score prot_model_0.cif pae_prot_model_0.npz --plddt plddt_prot_model_0.npz

Tenstorrent arithmetic differs from a GPU's, so the same inputs fold to slightly different numbers than they do on CUDA. See docs/interface-scores.md for the definitions, the choices made where the reference is ambiguous, and how this relates to the ipSAE BoltzGen reports while designing.

Affinity Predictions

For affinity targets, the same results.json entry also contains:

{
    "affinity_pred_value": 2.47,
    "affinity_probability_binary": 0.41,
    "affinity_pred_value1": 2.55,
    "affinity_pred_value2": 2.19,
    "affinity_probability_binary1": 0.50,
    "affinity_probability_binary2": 0.42
}

affinity_probability_binary: Probability of binding (0-1). Use for hit discovery (higher = more likely to bind)
affinity_pred_value: Predicted binding affinity as log10(IC50) in μM. Use for ligand optimization (lower = stronger binding). Only compare between known active molecules
affinity_pred_value1, affinity_pred_value2: Individual model predictions for binding affinity
affinity_probability_binary1, affinity_probability_binary2: Individual model predictions for binding probability
runtime_s: Wall-clock seconds for the whole target, structure and affinity together. Affinity targets also carry structure_runtime_s and affinity_runtime_s; the affinity leg is normally the larger of the two by several times, so read the split before pricing a screen

Advanced Usage

Input Format Details

Proteins with Custom MSA

- protein:
    id: A
    sequence: MVTPEGNVSLVDES...
    msa: ./path/to/msa.a3m

Proteins with Modifications

- protein:
    id: A
    sequence: MVTPEGNVSLVDES...
    modifications:
      - position: 5
        ccd: PTR  # Modified residue code

Ligands

- ligand:
    id: B
    smiles: 'CC1=CC=CC=C1'  # SMILES string
    # OR
    ccd: ATP                # CCD code

Constraints

Pocket and contact constraints are Boltz-2 only (they need a trained constraint embedder). A covalent bond to a ligand or a modified residue works on every structure model that takes the ligand. A bond between two standard residues (a disulfide) works on Boltz-2, ESMFold2, Protenix and OpenDDE; RF3 and the OpenFold3 family refuse it by name.

Pocket Constraints (binding site):

constraints:
  - pocket:
      binder: B              # Ligand chain
      contacts: [[A, 10], [A, 11], [A, 12]]  # Binding site residues
      max_distance: 6.0      # Angstroms (4-20A, default 6A)
      force: false           # Use potential to enforce (default: false)

Contact Constraints:

constraints:
  - contact:
      token1: [A, 10]
      token2: [A, 50]
      max_distance: 8.0
      force: false

Bond Constraints (covalent link, e.g. a covalent inhibitor, glycosylation, or disulfide):

constraints:
  - bond:
      atom1: [A, 10, SG]     # [chain, residue, atom]
      atom2: [B, 1, C1]      # SMILES ligand: element + count in SMILES order

OpenDDE + covalent bonds: OpenDDE honors a bond constraint between a protein residue and a ligand atom (the covalent-inhibitor case) or between two protein residues (a disulfide or crosslink). Both ride the same token_bonds machinery as Protenix-v2 and are honored in the output (device-verified against upstream OpenDDE within the reference's own seed noise floor); see examples/opendde_covalent_ligand.yaml and examples/opendde_covalent_bond.yaml.

Templates

Use experimental structures as templates:

templates:
  - cif: ./template.cif
    chain_id: A
    template_id: A
    force: true              # Enforce template alignment
    threshold: 2.0           # Max deviation in Angstroms

chain_id names the chains to template (default: every protein chain) and template_id the template's chains by mmCIF label_asym_id (default: the best match). Each chain is aligned to the template's sequence for you. This block works on boltz2, protenix-v2, opendde, opendde-abag, openfold3, openbind and rf3. force (with its threshold) and pdb files are Boltz-2 only; the other models refuse them, and RF3 takes one template per chain.

The same models except boltz2 also take a precomputed alignment .npz per protein chain (the format the upstream benchmark cache ships); Boltz-2 refuses it and takes the template as a cif instead. Its structures are fetched from RCSB, and a missing one is a hard error rather than a silently dropped template. See examples/7xi5_tmpl.yaml. There is no template search.

sequences:
  - protein:
      id: A
      sequence: MSSATPDPAEILT...
      templates: ./templates.npz

Command-Line Options

Model-specific options are labelled below.

Common Options:

Option	Default	Description
`--model`	`boltz2`	`boltz2`, `esmfold2`, `esmfold2-fast` (single-sequence ESMFold2; protein / DNA / RNA / ligand complexes), `protenix-v1` / `protenix-v2` (AlphaFold3-family folder; protein / RNA / DNA / ligand complexes; `protenix-v1` is upstream's v0.5.0 base checkpoint at 4 trunk recycles, `protenix-v2` the wider one at 10), `openfold3` (AlphaFold3-family folder; protein / RNA / DNA polymers, optional templates, `OF3_CKPT` weights), `openbind` (OpenBind-0; the OpenFold3 stack on upstream v0.5.0 weights, protein-ligand co-folding, `TT_BIO_OPENBIND` weights), `opendde` / `opendde-abag` (antibody-antigen co-folding on the Protenix-v2 stack plus a structural-token expander; `opendde-abag` selects the antibody-antigen checkpoint; protein-only for now), or `rf3` (RoseTTAFold3, AlphaFold3-family folder; protein / RNA / DNA / ligand complexes)
`--out_dir`	`./`	Output directory
`--cache`	`~/.boltz`	Weight cache directory. Whole-repo models (ESMFold2, ESMC, SaProt, OpenDDE) use the Hugging Face cache; `TT_BIO_CACHE` moves both, see docs/weights.md
`--accelerator`	`tenstorrent`	(Boltz-2) `tenstorrent`, `cpu`, or `gpu`; other models run on Tenstorrent
`--recycling_steps`	model-specific	3 for Boltz-2 and OpenFold3; 4 for Protenix-v1 (its checkpoint's own `N_cycle`); 10 for Protenix-v2/OpenDDE/ESMFold2/RF3 (the ESMFold2 paper's benchmark setting). Boltz-2, ESMFold2 and the OpenFold3 family run one more trunk cycle than asked and accept 0; Protenix, OpenDDE and RF3 count cycles, so their smallest value is 1
`--sampling_steps`	model-specific	Requested diffusion sampling steps: 200 for Boltz-2/Protenix-v1/Protenix-v2/OpenFold3/OpenDDE; 100 for ESMFold2 (executes 68 after the sigma-schedule clip, the paper's protocol)
`--diffusion_samples`	`1`	Number of structure samples. Device memory stays flat past the chunk width and time grows linearly, see docs/sample-scaling.md
`--partial_t`	`0`	rf3 only. Schedule index the diffusion rollout starts at, so it refines `--partial_structure` instead of folding from scratch. Higher stays closer to that structure
`--partial_structure`	—	rf3 only. The `.cif`/`.pdb`/`.json` structure `--partial_t` refines. It supplies the sequences too, so no MSA is attached
`--early_stop_plddt`	—	rf3 only. Abandon a target after the first trunk recycle if its mean pLDDT is below this. Writes no structure; the results entry carries `early_stopped`
`--max_parallel_samples`	`5`	(Boltz-2/Protenix/OpenDDE) Diffusion samples denoised in one batched forward. Device memory grows linearly in it; when the chip refuses a batch the fold halves it on its own, down to one sample, instead of failing. ESMFold2 sizes its own chunk to free memory; OpenFold3, OpenBind and RF3 denoise one sample at a time
`--output_format`	`cif`	`cif` or `pdb`. A PDB has one column for the chain id, so a longer name is rewritten `A`, `B`, `C`... and the originals go into a `REMARK 999` block; `cif` keeps them as submitted. See docs/model-capabilities.md
`--seed`	`0`	Random seed for the diffusion sampler
`--trace`	`False`	(Protenix-v1/Protenix-v2/OpenDDE) Replay a captured trace of the per-step diffusion device stream instead of dispatching it from the host every step. The output is identical to a run without it. On Wormhole at 512 tokens it did not change the end-to-end time, and it reserves 0.2-0.3 GB of device memory
`--diffusion_trace`	`False`	(Boltz-2) The same for Boltz-2's diffusion DiT stream; `tt-bio design --model boltzgen` takes the same flag
`--write_pde`	`False`	(Boltz-2) Write the PDE matrix to its own `<name>_pde.npz`. The Protenix family and OpenDDE put PDE next to PAE in one file under `--write_pae` instead
`--write_embeddings`	`False`	(Boltz-2) Write the `s`/`z` embeddings per target
`--override`	`False`	Re-run from scratch
`--use_msa_server`	auto	Use the online ColabFold API; auto-enabled for Boltz-2/Protenix-v1/Protenix-v2/OpenFold3/OpenBind-0/OpenDDE/RF3 when no local DB is found
`--single_sequence`	`False`	(Boltz-2/Protenix-v1/Protenix-v2/OpenFold3/OpenDDE) Skip all MSA requests; lower accuracy
`--msa_endpoint`	—	Fetch unpaired MSAs from a `tt-bio msa-server`. A complex is not paired through it unless its paired MSA is already in `--msa_dir`
`--write_pae`	`False`	(Protenix-v1/Protenix-v2/OpenDDE) Write the token-token PAE/PDE matrices to `<name>_pae.npz`
`--use_potentials`	`False`	(Boltz-2) Apply physical constraints
`--affinity_mw_correction`	`False`	(Boltz-2) Apply MW correction to affinity
`--num_devices`	`0`	Number of TT devices (0=all available)
`--device_ids`, `--devices`	—	Comma-separated TT device IDs (e.g. `0,2`); `--devices` is the shorter alias (matches `tt-bio embed`)
`--host_threads`	all cores	Total CPU threads this process may use, split across its cards. Set it when you run several single-card predicts side by side on one host: each one otherwise sizes its thread pools to every core and they fight for the CPU. Use cores ÷ concurrent predicts. At two threads per card or fewer the pools also stop spinning through device syncs (Tuning flags)
`--fast`	`False`	Makes some operations use a lower-precision numeric format that runs faster; accuracy is typically very close
`--report-energy`	`False`	(Boltz-2) Enables optional energy profiling for one TT device (requires `tt-mgmt` add-on); writes `power_profile.csv` and `power_profile.png`
`--energy-metric`	`both`	(Boltz-2) Choose power channel(s): `tdp`, `input`, or `both`
`--energy-sample-hz`	`20.0`	(Boltz-2) Sampling rate in Hz for both `power_w` and `input_power_w` channels

Affinity-Specific Options (Boltz-2):

Option	Default	Description
`--sampling_steps_affinity`	`200`	Sampling steps for affinity
`--diffusion_samples_affinity`	`5`	Number of affinity samples

MSA Options (Boltz-2, Protenix-v1, Protenix-v2, OpenFold3, OpenBind-0, OpenDDE and RF3 use an MSA by default; ESMFold2 only when requested):

Option	Default	Description
`--msa_db_path`	auto-detect	Path to local ColabFold database (`~/.boltz/msa_db` if present)
`--msa_dir`	`<out_dir>/msa`	MSA cache directory. Point it at a shared persistent path to reuse `{seq_hash}.a3m` across runs
`--msa_cache_only`	`False`	Treat `--msa_dir` as the only MSA source: never search, and fail rather than quietly fold a chain single-sequence
`--use_envdb`	`False`	Also search environmental database
`--use_msa_server`	auto	Use ColabFold API for MSA (auto-enabled when no local DB is found)
`--single_sequence`	`False`	Fold without an MSA (Boltz-2/Protenix-v1/Protenix-v2/OpenFold3/OpenDDE)
`--msa_server_url`	`https://api.colabfold.com`	MSA server URL
`--msa_pairing_strategy`	`greedy`	`greedy` or `complete`
`--max_msa_seqs`	`8192`	Maximum MSA depth. The default applies to Boltz-2 only. Unless you set it, the other models read what their upstream reads: ESMFold-2, Protenix, OpenDDE, OpenFold3 and OpenBind up to 16384 rows, RF3 1024 rows drawn per recycle. Each fold reports the depth it used as `msa_depth`
`--subsample_msa`	`False`	Subsample MSA
`--num_subsampled_msa`	`1024`	Number of subsampled sequences

MSA Database Setup Options:

Option	Default	Description
`--db`	`uniref30`	`uniref30` (~500GB), `envdb` (~800GB), or `all`
`--path`	`~/.boltz/msa_db`	Where to store the databases
`--install-tools`	`True`	Auto-install missing `mmseqs`/`colabfold_search`

Tuning Flags

The engine ships its device optimizations on. Each one is an environment variable you can set to 0 to fall back to the path it replaced, which is what you want if you are bisecting a result.

Flag	Default	What it does
`BOLTZ2_TOKEN_DIT_SDPA`	on	Boltz-2 only: runs the token-level DiT attention as one fused SDPA instead of materialising a 16x512x512 score matrix and reading it back. Not bit-exact — the kernel holds its exponentiated scores in bf16.
`TT_BIO_ATOM_AXIS_BUCKET`	on	Sizes the atom axis on the real atom count instead of assuming every token is a tryptophan. Byte-identical at 298 residues; at 512 it reassociates one matmul's contraction.
`TT_BIO_ATOM_SHIFT_GATHER`	on	Builds each atom's attention key window by slicing the atom sequence instead of selecting it with a matrix multiply. Same structure, bit for bit.
`TT_BIO_DEVICE_CONDITIONING`	on	Boltz-2 only: runs the diffusion conditioning's pair track on the card, where the trunk already left the tensor. Not bit-exact — device bf16 where the host path was fp32. Scored against 1HCL it is flat to slightly closer.
`TT_BIO_DEVICE_CONFIDENCE`	on	Boltz-2 only: assembles the confidence head's pair input on the card, where the trunk already left the tensor, instead of building it on the host and uploading it. Together with the flag below it folds 512 residues 1.0474x faster on Blackhole. Coordinates are bit-identical; only the confidence scores move (pLDDT by at most 0.362 of 100).
`TT_BIO_DEVICE_CONF_HEADS`	on	Boltz-2 only: runs the confidence head's pae/pde projections and their bin contractions on the card, so three numbers per token pair come down instead of a tile of bin logits: 67.1 MB becomes 2.1 MB at 512 residues on Wormhole, and 16.8 MB on Blackhole, where reading the tile whole and slicing on the host is 3x faster than the narrower row-major read. Coordinates and per-residue pLDDT are bit-identical; the confidence scores shift by a bf16 rounding, 0.0031 of pTM at 512 residues against a 0.0758 four-seed spread. Measured with the flag above.
`TT_BIO_DEVICE_ZINIT`	on	Boltz-2 only: builds the trunk's `z_init` pair tensor on the card instead of in torch, so the 134 MB upload at 512 tokens never happens. Folds 512 residues 1.01955x faster on Blackhole, ten of ten paired reps positive. Not bit-exact — device bf16 where the host path was fp32. Moves a 298 aa structure 0.264 Å all-atom inside its 0.35 Å bar, and native CA-lDDT against 1HCL is flat over eight seeds.
`TT_BIO_DIT_COND_HOIST`	on	Hoists the token diffusion transformer's conditioning out of the layer loop: every layer reads the same vector through six projections of its own, so folding each layer's norm scale into its own weight block replaces 144 matmuls and 48 layer norms per sampling step with one parameter-free norm and two concatenated matmuls. Same dot products, same FLOPs; the launches are grouped differently. RF3's token DiT builds the same block and inherits this default. Not bit-exact — one bf16 rounding changes order. Worth +0.2052 s of a 512-residue fold (14.588 s to 14.392 s at 1350 MHz, five interleaved reps, 95 % CI [+0.1561, +0.2543] against a +0.0324 s A/A floor).
`TT_BIO_FUSE_BIAS_STACKS`	on	Boltz-2 only: builds the diffusion conditioning's per-layer bias stack in one pass instead of one call per layer. Not bit-exact — moves a 298 aa structure 0.218 Å all-atom, inside its 0.35 Å bar.
`TT_BIO_FUSE_MASK_ADD`	on	Runs the gated-residual write-back as one `ttnn.addcmul` instead of a multiply then an add. Same structure, bit for bit.
`TT_BIO_FUSE_NORM_RESIDUAL`	on	Passes an add whose only consumer is a layer norm to the norm as its residual input instead of running the add first. Same structure, bit for bit.
`TT_BIO_FUSE_SCALE_ADD`	on	Runs attention's scale-then-bias as one `ttnn.addalpha` instead of a multiply then an add, for fp32 operands only. Bit-identical at every call shape.
`TT_BIO_GATE_GRANULARITY`	2	Tiles per DST acquire in the reblock-permute gate kernel. Same structure, bit for bit at every value; 2 is the setting that wins on Blackhole without losing much on Wormhole.
`TT_BIO_HOST_LEVERS`	on	Master switch for the host-side Boltz-2 levers (`TT_BIO_FUSE_BIAS_STACKS` and `TT_BIO_HOST_BLOCK_PAIRWISE`). Set it to `0` to take the host path for all of them at once.
`TT_BIO_MSA_LADDER`	on	Boltz-2 and BoltzGen: pads the MSA depth axis to the smallest of 64, 128, 256, 512, 1024 that holds the alignment instead of always to 1024, so a shallow search stops carrying rows that are not there. Not bit-exact — a shorter rung reassociates the same terms. Scored against 1HCL it is as accurate or closer.
`TT_BIO_OPM_LEGACY_LAYOUT`	off	Restores `OuterProductMean`'s old output stage. By default the MSA mean's `1/depth` scalar is applied to the per-row MSA tensor (2,097,152 B at 512 residues) instead of to the assembled pair rows it used to multiply (536,870,912 B), and the output projection runs as one matmul over the flattened token rows instead of once per row against a pinned core grid. Reaches every model that builds the shared `OuterProductMean`: the Boltz-2 and BoltzGen trunk, Protenix, OpenFold3's MSA embedder, RF3 and AF2. Not bit-exact — unpinned, the projection sums in a different order, one bf16 step. On the hinged 512-residue fixture the two arms differ by 0.29 to 1.59 A worst pseudo-domain across four seeds, against 1.09 to 1.42 A for the old path against its own seeds, and neither arm separates from the other against the crystal. Worth 0.1577 s of a 512-residue fold (14.3923 s to 14.2346 s at 1350 MHz, six interleaved reps against a 0.0339 s A/A floor).
`TT_BIO_PAIR_FFN_L1_FC1`	on	ESMFold2 only: keeps both halves of the pair transition's first matmul in L1, so the SiLU multiply that consumes them reads on chip instead of out of DRAM. Folds 512 residues 1.0879x faster on Blackhole. Same structure, bit for bit; it pays up to 512 residues and is inert above, where the block leaves it no room.
`TT_BIO_PAIR_INPLACE`	on	Where a pair tensor is too big to have a second copy beside it on the card (over 1.5 GiB on a 12 GiB Wormhole chip, one eighth of DRAM elsewhere), writes each block of a triangle attention, transition or triangle multiplication result back into the pair tensor on the card instead of assembling the result on the host. Same structure, bit for bit. Only OpenDDE's structural-token refiner reaches that size at 1536 residues or below; with `TT_BIO_TRIMUL_INPROJ_ROWBLOCK_NORM` it folds 1536 residues on one Wormhole chip in 2826 s instead of 4221 s. See Tuning flags.
`TT_BIO_PWA_BATCH_HEAD_WEIGHTS`	on	Computes every attention head's MSA row weights from one projection of the pair tensor instead of one projection per head. Same structure, bit for bit.
`TT_BIO_REBLOCK_PERMUTE_GATED`	on	Folds a triangle multiplication's chunk and its two sigmoid gates into the channel move that feeds them. Same structure, bit for bit.
`TT_BIO_RESIDUAL_L1`	on	Has the two Pairformer sub-layers whose residual update used to go to DRAM and come straight back write it to L1 instead. Same structure, bit for bit; the update stays in DRAM above 512 residues, where it no longer fits.
`TT_BIO_SDPA_ADD_GRANULARITY`	auto	Batches the fused SDPA kernel's running-sum/max and mask adds instead of doing them one tile at a time. Same structure, bit for bit at every granularity.
`TT_BIO_SDPA_FUSED_LARGE_S`	on	Runs triangle attention through the fused mask kernel above 1024 tokens, where the chunk ladder otherwise hands the call back to the stock attention. Worth 4.23x on the attention op at 1536 tokens and 1.1856x on a 1536-residue fold at the shipped 200 sampling steps (27.3 s saved). Set it to `0` for the stock ladder everywhere. Not bit-exact above 1024 tokens, where nothing reached this kernel before: it moves a 1536-residue structure 1.007 Å where a different seed moves it 36.6 Å, and pLDDT comes out 0.35 of 100 higher. At and below 1024 tokens the path is untouched, byte for byte.
`TT_BIO_SDPA_GRID_Q_CHUNK`	on	Sizes each attention's query chunk to the card's compute grid instead of a fixed cap, so a small attention fills the cores it has. Same structure, bit for bit.
`TT_BIO_SDPA_WIDE_K`	on	Lets triangle attention take a wider SDPA key chunk at the twenty padded token lengths whose shipped chunk does not divide them (288, 352, 416, 544, 608, 704, 736, 832, 864, 928, 992, 1056, 1088, 1184, 1216, 1248, 1312, 1376, 1472, 1504), where the fused kernel used to decline every call and hand the fold back to the stock op. Every model buckets to a multiple of 32, so any model that reaches this kernel can present all twenty; OpenFold3, ESMFold2 and RFD3 reach this path at no length and are untouched. It does nothing at a length the 256 cap already divides, which includes 512, 768 and 1024. Worth 1.27x-4.39x on the op wherever it fires (Blackhole p150a, 13x10 grid, arms interleaved, median of three blocks); the fold-level figure is not quoted here because the one stage arm that exists recorded no clock and no board, and `docs/sdpa-wide-k-parity.md` says so. Not bit-exact at those twenty lengths — a wider chunk changes the online-softmax reduction order, and this path is otherwise byte-reproducible at a fixed seed, so it moves a 686-residue structure 0.060-0.146 Å where a different seed moves it 3.69-7.28 Å, and pLDDT by 0.0001 against a seed-to-seed 0.0041. Every other length is byte for byte the old pick, and `0` restores it everywhere.
`TT_BIO_TOKEN_BUCKET`	on	The one global off switch for token bucketing. Every model's bucket answers to it, and the legacy per-model flags are ANDed with it, so `0` turns all of them off at once and folds the exact token count instead of a padded one. Off-lattice counts are slower and are what `docs/size-generality.md` asks for when checking a size claim is not an artifact of the lattice.
`TT_BIO_TRANSITION_L1_ROWS`	on	Blackhole only: sizes each transition's row block from the card's own L1 budget instead of the 16 rows tuned for Wormhole, so a 512-residue pair tensor takes 11 row blocks where it used to take 32. Applies to Boltz-2, BoltzGen and OpenFold3, whose pair track is 128 channels wide; Protenix-v2 and OpenDDE have wider pair tracks and keep today's height. Bit-identical output on the reference fixture at 298, 512, 768 and 1024 residues; on other shapes the block boundary moves the structure a little, 0.165 Å on the no-MSA prot leg whose arms already sit 7 Å from the fp32 reference. Folds 512 residues 1.023-1.035x faster on Blackhole, 768 residues 1.030x and 1024 residues 1.013x; it does nothing at and above 1536 residues, where the fixed height already fills the budget.
`TT_BIO_TRIATT_FUSED_QKVG`	on	Projects a triangle attention's query, key, value and gate in one pass over the pair tensor instead of two. Same structure, bit for bit.
`TT_BIO_TRIATT_FUSED_QKVGB`	on	Adds the pair-bias projection to that same pass, so the pair tensor is read once instead of three times. Same structure, bit for bit; chains of 32 residues or fewer keep the separate projection.
`TT_BIO_TRIATT_SDPA_HIFI_AB`	on for the OpenFold3 trunk, off elsewhere	Runs a triangle attention through the fused SDPA at HiFi4 with the reference reduction order, instead of the materialised fp32-softmax chain. OpenFold3 folds 512 residues 1.5123x faster with it (34.138 s to 22.574 s on Blackhole, 1.7297x at 640 residues), and it declines every call at 1088 residues, where the fold stays byte-identical to the old route. Not bit-exact below that: at 298 residues it moves a structure 2.58-8.01 A CA where a different seed moves it 5.23-9.78 A, and it lands 0.396 A closer to the deposited structure than the route it replaces. Per construction site, so a bare `openfold3.trunk` forces it on, `-openfold3.trunk` forces it off, and `all` / `-all` do the same to every site with no token of its own. Boltz-2 and RF3 build the same block and ship off.
`TT_BIO_TRIMUL_FUSED_GOUT`	on	Computes a triangle multiplication's output gate as a second output of its input projection. Same structure, bit for bit.
`TT_BIO_TRIMUL_GP_BANK_SPLIT`	on	Interleaves the gate and value columns of a triangle multiplication's fused input projection so the channel move reads each pair from two DRAM banks instead of one. Same structure, bit for bit; the gain is Blackhole's, free elsewhere.
`TT_BIO_TRIMUL_INPROJ_ROWBLOCK_NORM`	on	For a pair tensor over 3 GiB, computes a triangle multiplication's input projection in row blocks that each normalise their own rows, instead of joining each channel group's projection on the host. Same structure, bit for bit. Only OpenDDE's structural-token refiner reaches that size at 1536 residues or below. See Tuning flags.
`TT_BIO_TRIMUL_MASK_AFTER_MOVE`	on	Applies a triangle multiplication's pair mask after the channel move rather than before it. Same structure, bit for bit.
`TT_BIO_TRIMUL_MASK_L1`	on	Keeps a triangle multiplication's pair mask in L1, where the channel-blocked multiply re-reads it without touching DRAM. Same structure, bit for bit.
`TT_BIO_TRIMUL_MM_TRANSPOSE`	on	Lets the matmul take a triangle multiplication's operand transpose instead of running a separate transpose first. Same structure, bit for bit; `--fast` keeps the separate op, because transposing inside the matmul re-quantises a block-float tile.
`TT_BIO_TRIMUL_TAIL_F1`	on	Runs a triangle multiplication's output projection, gate projection and gate multiply as one kernel. Same structure, bit for bit.
`TT_BIO_TRIMUL_TAIL_F1_L1_OUT`	on	Lands that fused tail's product in L1 rather than DRAM, so the gate multiply that consumes it reads on chip. Same structure, bit for bit.
`TT_PROTENIX_CONF_DEVICE`	off	Protenix-v2 and OpenDDE: runs the confidence head on the card instead of the host. Off by default because pLDDT is precision-sensitive here; coordinates are unaffected either way.

More on how these were measured, and what "same structure" means for each of them, in docs/tuning-flags.md.

MSA Server Authentication

For --use_msa_server:

Basic Authentication:

export BOLTZ_MSA_USERNAME=myuser
export BOLTZ_MSA_PASSWORD=mypassword
tt-bio predict ... --model boltz2 --use_msa_server

API Key Authentication:

export MSA_API_KEY_VALUE=your-api-key
tt-bio predict ... --model boltz2 --use_msa_server

Running tt-bio on many machines

Each machine runs its own tt-bio controller, which listens on 127.0.0.1 and keeps a worker on every chip. Work reaches it through --controller, and whatever spreads jobs across machines sits on top: a platform, a cluster scheduler, or the fifty-line examples/many_hosts.py, which folds a directory across several hosts over ssh. docs/multi-host.md is the contract a scheduler builds against: the endpoints, what a host advertises, the lease and how a result settles once.

Optional: Energy Measurement (Boltz-2)

Use --report-energy to profile energy during prediction:

tt-bio predict examples/686.yaml --model boltz2 --override --device_ids 0 --report-energy --energy-metric both --energy-sample-hz 5

Behavior:

Select metric channel(s) with --energy-metric (tdp, input, both)
Uses one sampling rate (--energy-sample-hz, default 20 Hz)
Supports only Tenstorrent runs with one selected device
Records two power channels when available:
- power_w: tt-mgmt UMD telemetry power (TDP channel)
- input_power_w: tt-mgmt UMD telemetry input power
Requires optional tt-mgmt installation:
- git clone --recursive https://github.com/aperezvicente-TT/tt-mgmt.git
- pip install -e ./tt-mgmt
Prints energy summary metrics for selected channels
Always writes:
- power_profile.csv
- power_profile.png

Design

Design new binders and protein structures from a target or motif specification: one command, two models:

tt-bio design examples/binder.yaml --model boltzgen --num_designs 10
tt-bio design specs.json --model rfd3 --from_pdb --out_dir designs/

Model	Designs	Input
`boltzgen` (default)	protein / peptide / nanobody / antibody binders against a target	design YAML, same entity grammar as `predict`
`rfd3`	all-atom structures: binders, motif scaffolding, nucleic-acid binders	JSON spec with contig strings
`pxdesign`	binder backbones against a target structure	target YAML: structure file, chains to condition on, binder length

BoltzGen designs binders against a target structure. The pipeline runs design → inverse folding → folding → analysis → filtering and writes the top-ranked binders to <out_dir>/final_ranked_designs/. Pass --seed N to make a design reproducible; without it every run draws fresh. Input grammar, protocols, pipeline subsets, and options: docs/boltzgen-design.md. Designability (scRMSD) QA: docs/boltzgen-designability.md.

RFdiffusion3 (RFD3) is an all-atom generative model that designs new protein structures and sequences from a specification, rather than folding an existing one. Design modes, the contig-string input grammar, and which conditioning fields a spec can and cannot ask for: docs/rfd3-design.md.

PXDesign generates binder backbones against a target structure, conditioned on a distogram of the target rather than its coordinates. Input is a target YAML naming a structure file, the chains to condition on (with optional per-chain crop and hotspots) and a binder_length; each design is written as a CIF in the target structure's own frame, so it opens alongside your input file. A designs.json lands beside them with each design's numbers: fit RMSD against the target, binder residue and atom counts, and how many target tokens it was conditioned on. The binder is written as GLY because PXDesign generates a backbone with no sequence. Hotspot residues are label_seq numbers, not the author numbering a viewer shows, and a number that names no residue is refused rather than dropped. --num_designs is also the batch axis for this model: every requested design comes from one batched diffusion trajectory, and the gain per design grows with the batch and shrinks with the target: 2.7x at 8 designs against a 256-residue target, 1.5x against a 512-residue one, and flat from 16 up rather than turning back. A given --seed and --num_designs always reproduce the same designs, but --num_designs 1 and --num_designs 2 do not share their design 0: asking for more designs currently changes which ones you get, so pin both values when you want a run back. Selecting designs, which upstream does with a Protenix and an AF2-IG filter, is not on the CLI yet.

Each model downloads its weights automatically on first use. BoltzGen and RFdiffusion3 fan out across every available card (--devices 0,2 restricts); PXDesign runs on one card locally, or one design per card across a host's controller with --controller http://127.0.0.1:8765. tt-bio gen still works as a deprecated alias for tt-bio design --model boltzgen.

How many designs a card returns per hour, how --num_designs and --devices move it, and how to size a campaign: docs/design-throughput.md.

BindCraft 2 is not a tt-bio model and has no CLI entry; it is a third-party design loop you install yourself, and tt_bio.bindcraft2 gives it an AlphaFold 2 Evoformer that runs on a card. Its gradient loop runs on card, with the validation ensemble on BindCraft 2's own trunk so it stays the reference's; design acceptance is still being qualified. What you need, how to point a campaign at a chip, and what is not settled: docs/bindcraft2.md.

Training

Fine-tune or pre-train a model you can already run, with the same forward the inference path uses. The surface has four levels and you pick the one that matches what you want to write, not how much configuration you are willing to tolerate.

Level	You write	You own
`tt-bio finetune ...`	a command line	the config
`train.finetune(...)`	one call	the objective
`plan`, `batches`, `objectives`, `AdamW`, `Checkpointer`, `Mesh`, `trainable`	the loop	the `for` statement
`tt_bio.autograd` + `train.gradcheck`	an op and its backward	the gradient

Dropping a level is not a rewrite. train.recipes.source("default") prints the body the one-call version runs, written only in names the level below exports, and a test keeps it that way: if the recipe ever needed a private hook, the test fails and the hook becomes public.

Going wider or going deeper is one argument, at whichever level you are already on. These three are the same command:

# one chip: LoRA adapters on a frozen trunk
tt-bio finetune data/ --model protenix-v2 --out runs/a --global-batch 8 --steps 2000

# two chips: same run, one flag
tt-bio finetune data/ --model protenix-v2 --out runs/b --global-batch 8 --steps 2000 --chips 0,2

# pre-training: train the weights themselves, same loop
tt-bio finetune data/ --model protenix-v2 --out runs/c --global-batch 8 --steps 200000 \
    --train weights

The same three at the level below are train.finetune(...), plus mesh=, plus train="weights". Nothing is rewritten between them: one loop body serves both training modes and both chip counts, which the escape-hatch test checks instruction for instruction.

# will this fit, and how long? answered without opening a card
tt-bio finetune data/ --model protenix-v2 --out runs/a --global-batch 8 --steps 2000 --dry-run

tt-bio finetune --show-recipe        # the loop it would run, as source you can edit
tt-bio finetune --list-objectives    # the named loss rows

What works today: the interface, the dry run, both training modes, the optimizer, gradient checking, checkpoints, and data parallelism across the chips in one box. What does not: no model ships a training featuriser yet, so a real tt-bio finetune run stops with a named error at the point it would read your data. Featurisation is per model on purpose, and a model registers its own with tt_bio.train.catalogue.register. --train weights also comes back UNMEASURED from the dry run: we have measured a frozen trunk's memory and not a trained one's, and it will not print a projection shaped like a measurement.

finetune follows OpenFold3's optimizer setup rather than Adam's library defaults, which differ in three places that no loss curve shows: betas=(0.9, 0.95), no weight decay, and the AlphaFold 2 learning-rate schedule. Each is an argument, and the loop clips every sample separately, so a batch of 8 is 8 forwards per step. See docs/training.md for what each one costs if you get it wrong.

A training step runs softmax and layer norm in float64 on the host, which is what brings the OpenFold3 gradient inside its accuracy bar against upstream. It makes a step slower; --device-ops puts them back on the device kernels. Inference is unaffected. See docs/training.md.

Four things the API enforces rather than documents, because each is a bug we hit:

plan() answers from measured numbers or returns UNMEASURED. It refuses a crop size whose forward is measured to run out of memory instead of estimating one, and it will not report a 4-chip step time from a 2-chip measurement. Protenix's own 384-token crop is one of the refusals.
The optimizer refuses a bfloat16 master copy of the weights. An update accumulated at bfloat16 stops moving the weight while the gradient still looks healthy.
opt.step() raises if you spread training over several chips and never gave it a way to combine their gradients. Otherwise you train one model per chip and see one loss curve.
Every run records the clock it actually ran at, sampled during the work, plus the seed and the commit. A time without its clock is not a measurement on this hardware.

Global batch is always yours to set and is never derived from how many chips you have, so a recipe means the same thing on a bigger box.

Training on several chips is one flag, --chips 2, or one argument, mesh=train.Mesh({"dp": [0, 1]}). Under it a launcher runs your program once per chip and sums the gradients between them, the way torchrun does, so the program has to be re-runnable and must not open a card before the finetune call. Both are checked before anything starts. Two p150a chips on a QuietBox measured 1.96x at a 0.33 MB adapter gradient and 1.70x at 5.24 MB, both at 1350 MHz; the gap is host-side Adam contending between the two processes, not the exchange, which costs 2.3 % of the step. Four chips runs the same path and is not measured yet. One host: reaching a second box needs a cable, not a code change.

Every run carries the check that makes a multi-chip number mean something. The ranks' weights must stay identical, so the launcher compares every rank's master weights at the end and refuses a run where they differ, and it reads each rank's chip off that rank's own open file descriptors rather than trusting TT_VISIBLE_DEVICES, which names a different number than the device node. Both failures look like a healthy run otherwise.

Tiers, cut lines, the escape-hatch test and where plan() gets its numbers: docs/training.md. What one gradient step through a pairformer stack costs, how it scales with depth and width, and where the headroom is: docs/gradient-step-cost.md.

Cite

If you use this code or the models in your research, please cite the following papers:

@article{passaro2025boltz2,
  author = {Passaro, Saro and Corso, Gabriele and Wohlwend, Jeremy and Reveiz, Mateo and Thaler, Stephan and Somnath, Vignesh Ram and Getz, Noah and Portnoi, Tally and Roy, Julien and Stark, Hannes and Kwabi-Addo, David and Beaini, Dominique and Jaakkola, Tommi and Barzilay, Regina},
  title = {Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction},
  year = {2025},
  doi = {10.1101/2025.06.14.659707},
  journal = {bioRxiv}
}

@article{stark2025boltzgen,
  author = {Stark, Hannes and Faltings, Felix and Choi, MinGyu and Xie, Yuxin and Hur, Eunsu and O'Donnell, Timothy John and Bushuiev, Anton and U{\c c}ar, Talip and Passaro, Saro and Mao, Weian and Reveiz, Mateo and Bushuiev, Roman and Pluskal, Tom{\'a}{\v s} and Sivic, Josef and Kreis, Karsten and Vahdat, Arash and Ray, Shamayeeta and Goldstein, Jonathan T. and Savinov, Andrew and Hambalek, Jacob A. and Gupta, Anshika and Taquiri-Diaz, Diego A. and Zhang, Yaotian and Hatstat, A. Katherine and Arada, Angelika and Kim, Nam Hyeong and Tackie-Yarboi, Ethel and Boselli, Dylan and Schnaider, Lee and Liu, Chang C. and Li, Gene-Wei and Hnisz, Denes and Sabatini, David M. and DeGrado, William F. and Wohlwend, Jeremy and Corso, Gabriele and Barzilay, Regina and Jaakkola, Tommi},
  title = {BoltzGen: Toward Universal Binder Design},
  year = {2025},
  doi = {10.1101/2025.11.20.689494},
  journal = {bioRxiv}
}

@article{wohlwend2024boltz1,
  author = {Wohlwend, Jeremy and Corso, Gabriele and Passaro, Saro and Getz, Noah and Reveiz, Mateo and Leidal, Ken and Swiderski, Wojtek and Atkinson, Liam and Portnoi, Tally and Chinn, Itamar and Silterra, Jacob and Jaakkola, Tommi and Barzilay, Regina},
  title = {Boltz-1: Democratizing Biomolecular Interaction Modeling},
  year = {2024},
  doi = {10.1101/2024.11.19.624167},
  journal = {bioRxiv}
}

@misc{candido2026language,
  author = {Candido, Salvatore and Hayes, Thomas and Derry, Alexander and Rao, Roshan and Lin, Zeming and Verkuil, Robert and others},
  title = {Language Modeling Materializes a World Model of Protein Biology},
  year = {2026},
  url = {https://biohub.ai/papers/esm_protein.pdf},
  note = {Preprint; ESMC / ESMFold2}
}

@misc{protenix2025,
  author = {{ByteDance AML AI4Science Team}},
  title = {Protenix: An AlphaFold3 Reproduction for Biomolecular Structure Prediction},
  year = {2025},
  url = {https://github.com/bytedance/Protenix}
}

@misc{openfold3,
  author = {{OpenFold Consortium}},
  title = {OpenFold3: An Open-Source Reproduction of AlphaFold3},
  year = {2026},
  url = {https://github.com/aqlaboratory/openfold-3}
}

@article{butcher2025rfdiffusion3,
  author = {Butcher, Jasper and Krishna, Rohith and Mitra, Raktim and Brent, Rafael Isaac and Li, Yanjing and Corley, Nathaniel and Kim, Paul T and Funk, Jonathan and Mathis, Simon Valentin and Salike, Saman and Muraishi, Aiko and Eisenach, Helen and Thompson, Tuscan Rock and Chen, Jie and Politanska, Yuliya and Sehgal, Enisha and Coventry, Brian and Zhang, Odin and Qiang, Bo and Didi, Kieran and Kazman, Maxwell and DiMaio, Frank and Baker, David},
  title = {De novo Design of All-atom Biomolecular Interactions with RFdiffusion3},
  year = {2025},
  doi = {10.1101/2025.09.18.676967},
  journal = {bioRxiv}
}

In addition if you use the automatic MSA generation, please cite:

@article{mirdita2022colabfold,
  title={ColabFold: making protein folding accessible to all},
  author={Mirdita, Milot and Sch{\"u}tze, Konstantin and Moriwaki, Yoshitaka and Heo, Lim and Ovchinnikov, Sergey and Steinegger, Martin},
  journal={Nature methods},
  year={2022}
}

License

tt-bio is released under the MIT License (see LICENSE) and is built on the MIT-licensed Boltz-2 / Boltz-1 code. It bundles third-party code, each under its upstream license: the ESMFold2 host-side reference under tt_bio/_vendor/ (the esm pipeline, MIT, © Chan Zuckerberg Biohub; and the HuggingFace ESMFold2 model definition, Apache-2.0), the OpenFold3 host-side data pipeline under tt_bio/_vendor/openfold3/ (Apache-2.0, OpenFold Consortium), the BoltzGen binder-design source under tt_bio/boltzgen/ (MIT, © Hannes Stärk), and RF3's host featurizer under tt_bio/_vendor/rf3/, tt_bio/_vendor/foundry/ and tt_bio/_vendor/atomworks/ (BSD-3-Clause, University of Washington / Institute for Protein Design). Protenix-v2, OpenFold3's on-device model, RFdiffusion3, RF3's on-device model, and PXDesign's are independent ttnn reimplementations (no upstream compute code is vendored); Protenix-v2's weights download from ByteDance's Hugging Face mirror under Apache-2.0, RFdiffusion3's and RF3's checkpoints download directly from the Institute for Protein Design (BSD-3-Clause), PXDesign's generator checkpoint downloads from ByteDance's release host under Apache-2.0 and its AF2-IG selection stage reads DeepMind's AlphaFold2 monomer pTM parameters (CC BY 4.0, from storage.googleapis.com/alphafold/), and OpenFold3's of3-p2-155k.pt and OpenBind-0's of3-ob-2025-06-30-174k.pt are the consortium's ungated public parameter releases, which you fetch yourself (the project is Apache-2.0, stated by upstream as free for academic and commercial use; the consortium publishes no separate parameter license). See NOTICE for sources, versions, and modifications.

Name		Name	Last commit message	Last commit date
.github/workflows		.github/workflows
docs		docs
examples		examples
perf		perf
scripts		scripts
site-redirect		site-redirect
site		site
tests		tests
tt_bio		tt_bio
.gitignore		.gitignore
CHANGELOG.md		CHANGELOG.md
LICENSE		LICENSE
MANIFEST.in		MANIFEST.in
NOTICE		NOTICE
README.md		README.md
RELEASING.md		RELEASING.md
pyproject.toml		pyproject.toml

Latest commit

History

Folders and files

Repository files navigation