Go modules · #2774 by repository stars
github.com/gpustack/gguf-parser-go
v0.26.0gpustack/gguf-parser-go ↗Last release 9 days ago
19 Aug 2026
Release timing varies
gaps range from 8 days to 5 months
Rarely documented
notes for 10 of the last 60 stable releases
Nothing withdrawn
no release was ever pulled
2 years old
135 releases · first in 2024
Release timeline
135 releases since 2024Releases
- v0.26.019 Aug 2026
Release notes
Open source →fix: resolve the sliding window layout per layer, mirror llama.cpp's …
…per-architecture periods
The sliding window pattern was a five-entry switch (llama4, phi3, gemma2,
gemma3, cohere2) while llama.cpp interleaves windows on 21 architectures.
Every architecture missing from the switch was charged a full, context-scaled
KV cache on all blocks: gpt-oss-20b about 2x (6.0 GiB instead of about 3.0 GiB
of KV at 131072 context), gemma-3n-E4B about 6.5x.Read <architecture>.attention.sliding_window_pattern when the file declares it
(a period as a scalar, or the per-layer use as a bool array, as Gemma 4 writes),
and fall back to a per-architecture period table mirrored from llama.cpp's
src/models/*.cpp, including the dense-first phasing the old modulo could not
express. The resolved layout lives in AttentionSlidingWindowLayers and the
estimator charges each layer by its own kind.Also honour the layers that hold no cache of their own: the layers behind
n_layer_kv_from_start reuse leading caches (gemma3n hardcodes 20, Gemma 4
derives it from attention.shared_kv_layers), and the NextN/MTP blocks of the
hybrid family are filtered out of the main context cache. Read
attention.key_length_swa / value_length_swa so windowed layers size their
heads separately (Gemma 4). - v0.25.013 Jul 2026
- v0.24.127 May 2026
Release notes
Open source →fix: integer overflow in tensor elements/bytes calculation
Signed-off-by: thxCode <[email protected]>
- v0.24.1-0.20260325020219-35c4501b75f725 Mar 2026pre-release
Nothing published for this version
- v0.24.011 Feb 2026
- v0.23.16 Jan 2026
- v0.23.06 Jan 2026
Release notes
Open source →fix: --token now works with --url, add --header for custom HTTP heade…
…rs (#16)
* fix: --token now works with --url, add --header for custom HTTP headers
* update readme
- v0.22.118 Aug 2025
- v0.22.1-0.20250815014413-0f24c9422ada15 Aug 2025pre-release
Nothing published for this version
- v0.22.1-0.20250812041445-437e967adf2b12 Aug 2025pre-release
Nothing published for this version
- v0.22.06 Aug 2025
- v0.21.425 Jul 2025
Release notes
Open source →fix: embedding model estimation
Signed-off-by: thxCode <[email protected]>
- v0.21.324 Jul 2025
- v0.21.218 Jul 2025
Nothing published for this version
- v0.21.115 Jul 2025
Nothing published for this version
- v0.21.011 Jul 2025
Nothing published for this version
- v0.20.510 Jul 2025
Nothing published for this version
- v0.20.47 Jul 2025
Nothing published for this version
- v0.20.37 Jul 2025
Nothing published for this version
- v0.20.24 Jul 2025
Nothing published for this version
- v0.20.2-0.20250702062826-7f916a398f992 Jul 2025pre-release
Nothing published for this version
- v0.20.11 Jul 2025
Nothing published for this version
- v0.20.01 Jul 2025
Nothing published for this version
- v0.19.118 Jun 2025
Nothing published for this version
- v0.19.1-0.20250604140319-c63a186d12f04 Jun 2025pre-release
Nothing published for this version
- v0.19.1-0.20250602025111-333741c7f51c2 Jun 2025pre-release
Nothing published for this version
- v0.19.030 May 2025
Nothing published for this version
- v0.18.2-0.20250527070649-63791fa40a2a27 May 2025pre-release
Nothing published for this version
- v0.18.127 May 2025
Nothing published for this version
- v0.18.023 May 2025
Nothing published for this version
- v0.17.621 May 2025
Nothing published for this version
- v0.17.515 May 2025
Nothing published for this version
- v0.17.412 May 2025
Nothing published for this version
- v0.17.39 May 2025
Nothing published for this version
- v0.17.23 May 2025
Nothing published for this version
- v0.17.13 May 2025
Nothing published for this version
- v0.17.029 Apr 2025
Nothing published for this version
- v0.16.025 Apr 2025
Nothing published for this version
- v0.15.024 Apr 2025
Nothing published for this version
- v0.14.118 Apr 2025
Nothing published for this version
- v0.14.013 Apr 2025
Nothing published for this version
- v0.13.2014 Mar 2025
Nothing published for this version
- v0.13.194 Mar 2025
Nothing published for this version
- v0.13.1827 Feb 2025
Nothing published for this version
- v0.13.1725 Feb 2025
Nothing published for this version
- v0.13.1612 Feb 2025
Nothing published for this version
- v0.13.1511 Feb 2025
Nothing published for this version
- v0.13.1410 Feb 2025
Nothing published for this version
- v0.13.1329 Jan 2025
Nothing published for this version
- v0.13.1228 Jan 2025
Nothing published for this version
- v0.13.1116 Jan 2025
Nothing published for this version
- v0.13.1015 Jan 2025
Nothing published for this version
- v0.13.10-0.20250114085710-c626beafc12214 Jan 2025pre-release
Nothing published for this version
- v0.13.914 Jan 2025
Nothing published for this version
- v0.13.85 Jan 2025
Nothing published for this version
- v0.13.73 Jan 2025
Nothing published for this version
- v0.13.628 Dec 2024
Nothing published for this version
- v0.13.512 Dec 2024
Nothing published for this version
- v0.13.49 Dec 2024
Nothing published for this version
- v0.13.4-0.20241209040953-0634e74b49289 Dec 2024pre-release
Nothing published for this version