-
-
Notifications
You must be signed in to change notification settings - Fork 6.2k
Pull requests: unslothai/unsloth
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
fix(studio): keep MTP-flavor GGUF siblings as separate variants (#7460)
#7463
opened Jul 26, 2026 by
Souravrajvi0
Contributor
Loading…
Studio: use the UI font scale tokens in the Agents settings tab
#7462
opened Jul 26, 2026 by
danielhanchen
Member
Loading…
tests: import bitsandbytes before the CUDA/ROCm spoof so Repo tests (CPU) is green again
#7461
opened Jul 26, 2026 by
danielhanchen
Member
Loading…
feat(studio): run chats in parallel in the Chat tab
#7455
opened Jul 26, 2026 by
danielhanchen
Member
Loading…
Studio: fix the "No model loaded" error, and optionally auto-download a model named in an API request
#7454
opened Jul 26, 2026 by
danielhanchen
Member
Loading…
fix(studio): refresh token count after model load (#7450)
#7453
opened Jul 26, 2026 by
Souravrajvi0
Contributor
Loading…
fix(rocm): prefer system LLVM runtime on native Linux
#7448
opened Jul 26, 2026 by
Oxygen56
Loading…
feat(studio): adjustable llama-server parallel slots from the web UI
#7447
opened Jul 25, 2026 by
knamroud
Loading…
Enhance checkpoint management and multi-dataset training support
#7445
opened Jul 25, 2026 by
iam-knoit
Loading…
studio/tests: cover the GGUF load ordering behaviourally and make the structlog stub order-independent
#7442
opened Jul 25, 2026 by
danielhanchen
Member
Loading…
docs(studio): fix stale gfx110X example in ROCR masking comments
#7440
opened Jul 25, 2026 by
LeoBorcherding
Collaborator
Loading…
tests: read checked-in files as UTF-8 instead of the platform default
#7438
opened Jul 25, 2026 by
LeoBorcherding
Collaborator
Loading…
unsloth start: keep the local subagent unattended and out of plan mode
#7437
opened Jul 25, 2026 by
danielhanchen
Member
Loading…
Studio: admission control on /v1/messages, slot pool that tracks --parallel
#7436
opened Jul 25, 2026 by
danielhanchen
Member
Loading…
install.sh: do not assume sudo consent when there is no terminal
#7435
opened Jul 25, 2026 by
LeoBorcherding
Collaborator
Loading…
Studio UI font-scale test: normalise paths so the allowlists work on Windows
#7434
opened Jul 25, 2026 by
LeoBorcherding
Collaborator
Loading…
Studio: make the startup public-IP lookup opt-in
#7433
opened Jul 25, 2026 by
LeoBorcherding
Collaborator
Loading…
Show release notes in the update popup, sourced from CHANGELOG.md
#7432
opened Jul 25, 2026 by
shimmyshimmer
Member
Loading…
Bypass fast_generate for flash_attention_2 models (StaticCache + FA2 produces gibberish)
#7429
opened Jul 24, 2026 by
pwasiewi
Loading…
Studio: fetch bare hostnames as https instead of refusing them
#7427
opened Jul 24, 2026 by
NilayYadav
Collaborator
Loading…
Studio: split parallel tool calls for Llama 3.x chat templates
#7426
opened Jul 24, 2026 by
NilayYadav
Collaborator
Loading…
Studio: keep automatic model loading toast visible until completion
#7425
opened Jul 24, 2026 by
oobabooga
Member
Loading…
Studio: recover tool-enabled GGUF chats after llama-server exits
#7424
opened Jul 24, 2026 by
oobabooga
Member
Loading…
Parse a .json dataset file as one JSON document instead of line-by-line
#7422
opened Jul 24, 2026 by
vineethsaivs
Contributor
Loading…
Studio: Expose GPU memory mode in unsloth run and unsloth start
#7421
opened Jul 24, 2026 by
oobabooga
Member
Loading…
Previous Next
ProTip!
Mix and match filters to narrow down what you’re looking for.