F
v1.2.30
1.2.30 (2026-10-21) Full Changelog: v1.2.29...v1.2.30 Bug Fixes training-sdk: retain adapter-rejected TITO artifacts through the sidecar's configured private trace store and shared retention limits
F
v1.2.29
1.2.29 (2026-10-21) Full Changelog: v1.2.28...v1.2.29 Features training-sdk: let a serverless FiretitanServiceClient sample a base model or another run's saved checkpoints without creating a model and without using a concurrent-training-run slot, via standalone sampler routes training-sdk: make FiretitanSamplingClient picklable across worker processes (connection state, shared-storage R3 binding,…
F
v1.2.28
1.2.28 (2026-10-20) Full Changelog: v1.2.27...v1.2.28 Features training-sdk: add FiretitanServiceClient.delete_training_session() and FireworksClient.delete_training_session(name) to close a serverless training session and release its concurrent-run quota slot immediately instead of waiting for idle expiry training-sdk: capture complete replay masks independently of optional SC heads, preserving m…
F
v1.2.27
1.2.27 (2026-10-17) Full Changelog: v1.2.26...v1.2.27 Features training-sdk: accept keyword-only images= on FiretitanSamplingClient.sample for preprojected vision prompts, and reject unprojected image chunks so sampler-owned delimiters are not duplicated Bug Fixes training-sdk: remove the unused TITO parser fallback, distinguish upstream failures from local parsing failures, and retain bounded loc…
F
v1.2.26
1.2.26 (2026-10-15) Full Changelog: v1.2.25...v1.2.26 Features training-sdk: accept the built-in REINFORCE loss and expose trainer capabilities for REINFORCE, optional score centering, and shared server policy TIS/MIS; unsupported images fail before forward or backward submission
F
v1.2.25
1.2.25 (2026-10-15) Full Changelog: v1.2.24...v1.2.25 Bug Fixes training-sdk: overlap forward/backward request uploads across calls while retaining upload limits, retry ordering, and optimizer admission barriers; safely retire shared admission fences before multi-chunk forward calls
F
v1.2.24
1.2.24 (2026-10-13) Full Changelog: v1.2.23...v1.2.24 Bug Fixes training-sdk: preserve Router Replay fallback when validation errors omit rejected input values training-sdk: preserve configured routing headers when resolving the account for sampling Refactors training-sdk: keep training regression tests outside the installed SDK package while retaining them in source distributions
F
v1.2.23
1.2.23 (2026-10-12) Full Changelog: v1.2.22...v1.2.23 Features training-sdk: record each completed sampling attempt's client-measured latency and generated-token count in ServerMetrics, so callers can compare client and serving time from DeploymentSampler.drain_metrics() training-sdk: DeploymentSampler.drain_metrics() now returns one record per completed or failed attempt: every record carries its…
F
v1.2.22
1.2.22 (2026-10-10) Full Changelog: v1.2.21...v1.2.22 Features training-sdk: preserve optional Parquet sampling-support references through deployment sampling, Tinker-style sampled sequences, model inputs, and TITO; target-policy support selection remains off by default training-sdk: accept the trainer's built-in DPPO loss name Bug Fixes training-sdk: retain HTTP request/response diagnostics when…
F
v1.2.21
1.2.21 (2026-10-08) Full Changelog: v1.2.20...v1.2.21 Features training-sdk: send projection_head_dim as first-class trainer configuration so any compatible validated training shape can create the requested head, while preserving legacy extra_args
F
v1.2.20
1.2.20 (2026-10-06) Full Changelog: v1.2.19...v1.2.20 Features training-sdk: allow TITOSidecar(bind_address=...) to listen on a specific host interface, including Docker bridge gateways; loopback remains the default Bug Fixes dependencies: allow base SDK installs with Pydantic 2.13 while retaining the Pydantic <3 upper bound. Training extras retain their existing Pydantic <2.13 constraint. depende…
F
v1.2.19
1.2.19 (2026-10-02) Full Changelog: v1.2.18...v1.2.19 Features training-sdk: automatically select RDMA weight sync only for dedicated full-parameter trainers and inference replicas that all advertise support; older images, LoRA, and deployments using gateway shard fan-out retain save-weights followed by file hotload datasets: add fireworks.lib.dataset_upload.upload_dataset_shards to upload a .json…
F
v1.2.18
1.2.18 (2026-10-01) Full Changelog: v1.2.17...v1.2.18 Features training-sdk: add forward-only reads for per-token projection outputs training-sdk: expose supports_router_replay on training clients, reported by the trainer at create_model, so callers can decide Router Replay without fetching the base model (None when the trainer does not report it) training-sdk: raise ModelDetailsUnavailableError (…
F
v1.2.17
1.2.17 (2026-09-26) Full Changelog: v1.2.16...v1.2.17 Features training-sdk: add lora_init_method to create_training_client (init_method on create_lora_training_client) to pick each model's LoRA initialization at create time: "kaiming" (default) or "nora" (NoRA-init)
F
v1.2.16
1.2.16 (2026-09-25) Full Changelog: v1.2.15...v1.2.16 Features training-sdk: add projection_head_dim to managed service and training-client creation for train-only projection heads, and output="projection" for forward_backward_custom gradients with respect to projection outputs
F
v1.2.15
1.2.15 (2026-09-22) Full Changelog: v1.2.14...v1.2.15 Features training-sdk: retain per-token top-logprob token IDs and values in sampling results and TITO trajectory artifacts Bug Fixes training-sdk: preserve multidimensional custom-loss gradients as shaped tensor data during backward requests training-sdk: preserve failed trainer status when reporting provisioning errors to managed training jobs…
F
v1.2.14
1.2.14 (2026-09-18) Full Changelog: v1.2.13...v1.2.14 Features training-sdk: select Parquet R3 using a small completion request; older inference servers retain inline routing without a capability endpoint training-sdk: allow sampling requests to set r3_ttl_seconds for shared Parquet routing files; the default retention is six hours
F
v1.2.13
1.2.13 (2026-09-18) Full Changelog: v1.2.12...v1.2.13 Features deployments: add accept_shapeless_risk opt-out on deployments.create for intentionally creating a deployment without a pre-validated shape; document the deployment_shape="default" sentinel that asks the server to pick a default validated shape training-sdk: add accept_shapeless_risk to DeploymentConfig for the manual shapeless creation…
F
v1.2.12
1.2.12 (2026-09-15) Full Changelog: v1.2.11...v1.2.12 Bug Fixes training-sdk: stop sending the obsolete managed-rollout deployment annotation training-sdk: wait for control-plane readiness when reattaching rollout deployments training-sdk: avoid copying dense tensor values when sizing requests for every client training-sdk: automatically negotiate optimized training requests for dedicated and serv…
F
v1.2.11
1.2.11 (2026-08-31) Full Changelog: v1.2.10...v1.2.11 Features training-sdk: add environment-local TITO sidecars with full-history prompt construction by default and an experimental incremental renderer protocol training-sdk: preserve and attribute completion serving metrics when merging response headers with final performance metrics training: support explicit reservation targeting for dedicated…
F
v1.2.10
1.2.10 (2026-08-30) Full Changelog: v1.2.9...v1.2.10 Features training-sdk: expose merged-base output precision with source-format export as the default Bug Fixes training-sdk: fail trainer readiness waits fast when the job stops instead of polling to the readiness timeout training-sdk: preserve canonical structured status details for managed recipe failures training-sdk: preserve Lifecycle, Tinke…
F
v1.2.9
1.2.9 (2026-08-12) Full Changelog: v1.2.8...v1.2.9 Bug Fixes training-sdk: preserve per-request affinity for serverless training sampling
F
v1.2.8
1.2.8 (2026-08-12) Full Changelog: v1.2.7...v1.2.8 Features training: default managed SFT/DPO and dedicated Training API trainers to reservation-first placement with an explicit opt-out
F
v1.2.7
1.2.7 (2026-08-07) Full Changelog: v1.2.6...v1.2.7 Features training: support reservation-first placement for managed SFT/DPO and dedicated Training API trainers training-sdk: expose managed SFT and DPO renderer Hugging Face repo ID selection
F
v1.2.6
1.2.6 (2026-08-03) Full Changelog: v1.2.5...v1.2.6 Features training: support CMEK-encrypted managed RFT sampler hot-loads and isolate reference trainers training-sdk: support repeated per-adapter model creation on one managed trainer Bug Fixes release: avoid false promotion-source failures after successful SDK releases
F
v1.2.5
1.2.5 (2026-07-31) Full Changelog: v1.2.4...v1.2.5 Bug Fixes training-sdk: log backend training shape override warnings
F
v1.2.4
1.2.4 (2026-07-27) Full Changelog: v1.2.3...v1.2.4 Features training-sdk: expose MoE model capability lookup for Router Replay training-sdk: expose structured Training API errors training-sdk: propagate training skill attribution headers Bug Fixes training-sdk: gate RLOR inactivity cleanup compatibility to Fireworks-owned jobs
F
v1.2.3
1.2.3 (2026-07-21) Full Changelog: v1.2.2...v1.2.3 CI require generated promotion provenance before public SDK releases
F
v1.2.2
1.2.2 (2026-07-21) Full Changelog: v1.2.1...v1.2.2
F
v1.2.1
1.2.1 (2026-07-20) Full Changelog: v1.2.0...v1.2.1 Features serverless-rl: single retry owner + structured sampling error (SDK) (7113d1a) training-sdk: adjust rollout concurrency within RL steps (68595ff) training-sdk: expose routing matrices in sample responses (81d52c2) training: serverless cross-run resume — SDK config-from-checkpoint + cookbook unlock (PR3+PR5) (a284467) Bug Fixes training-sdk…