Commit Graph
14207 Commits
Author SHA1 Message Date
geohot f26128e8d2 400 lines 2026-08-03 16:28:20 +00:00
geohot 79ef0a16c0 +499 lines 2026-08-03 15:02:58 +00:00
geohot 0dda620216 llm: fix recurrent rangeify and client cache reuse 2026-08-03 03:30:32 +00:00
geohot 6414b1b2e1 Merge remote-tracking branch 'origin/master' into qwen36_27b_amd_900 2026-08-03 03:03:46 +00:00
chenyuandGitHub 23c7813f44 no hard coded dtype int for rangeify debuf [PR] (#17373) 2026-08-02 20:43:04 -04:00
geohot de37fe0d8f kernel cleanups 2026-08-02 22:00:18 +00:00
geohot 269f821421 llm: retain master model formatting 2026-08-02 20:25:41 +00:00
geohot 41b7a02103 Merge remote-tracking branch 'origin/master' into qwen36_27b_amd_900
# Conflicts:
#	tinygrad/llm/model.py
2026-08-02 20:12:15 +00:00
geohot 9230322266 llm: keep sequential AMD startup under a minute 2026-08-02 20:04:55 +00:00
George HotzandGitHub 05bc7c6994 fix llm reasoning and Linear import (#17372) 2026-08-02 12:31:52 -07:00
chenyuandGitHub 09dabfe05e minor pm_float_decomp cleanup [PR] (#17371)
make the rule order independently correct
2026-08-02 15:17:33 -04:00
geohot ffc4507fb4 llm: recover packed weights in linear 2026-08-02 19:06:45 +00:00
chenyuandGitHub 59df317b12 use UOp.const to create new consts [PR] (#17368)
replace arg won't work with ConstArg
2026-08-02 13:31:55 -04:00
geohot 71869224b4 llm: share AMD quant kernel helpers 2026-08-02 15:53:39 +00:00
geohot c64f8aeff8 Merge remote-tracking branch 'origin/master' into qwen36_27b_amd_900 2026-08-02 15:38:58 +00:00
geohot 82025dfe1f llm: keep 131k Qwen cache on AMD 2026-08-02 15:38:04 +00:00
geohot 756a936be2 llm: report interrupted stream statistics 2026-08-02 15:12:29 +00:00
geohot ce071d7f0e llm: fix recurrent server startup and reasoning routing 2026-08-02 15:03:17 +00:00
geohot 21f91bac40 llm: reuse generic Qwen embedding 2026-08-02 14:44:24 +00:00
geohot 73f0aa7d45 llm: add Qwen AMD acceptance benchmark 2026-08-02 14:35:04 +00:00
geohot d1e1d587af llm: reuse standard sampling path 2026-08-02 14:32:51 +00:00
geohot 8f3337c46f llm: checkpoint recurrent prompt state 2026-08-02 14:30:06 +00:00
geohot f92ac85fbd llm: simplify AMD Qwen kernels 2026-08-02 14:17:43 +00:00
geohot fb53442c17 llm: use standard RoPE table allocation 2026-08-02 14:13:06 +00:00
geohot eba359128e llm: reuse master recurrent cache semantics 2026-08-02 14:09:36 +00:00
geohot 3804a87b60 llm: keep Qwen warmup changes scoped 2026-08-02 13:55:10 +00:00
geohot 9d02a3fe68 llm: remove unused Qwen AMD paths 2026-08-02 13:55:10 +00:00
geohot 35dbe53d5d llm: trim unused AMD quant paths 2026-08-02 13:55:10 +00:00
geohot ca29776f74 llm: fast packed Qwen3.6 inference on AMD 2026-08-02 13:55:10 +00:00
wozeparrotandGitHub e14cadb1fb gptoss: set ASM_GEMM (#17363) 2026-08-02 06:52:36 -07:00
chenyuandGitHub 0258c7fefc minor argstr and alloc stack cleanup [PR] (#17361) 2026-08-01 22:37:06 -04:00
George HotzandGitHub fb607fb990 faster devectorizer with one line (#17360) 2026-08-01 13:06:35 -07:00
George HotzandGitHub 15c936db01 merge devectorize + indexing (#17354)
* external benchmark schedule in 5 sec (codex slop)

* prune

* real?

* delete
2026-08-01 11:40:15 -07:00
wozeparrotandGitHub 98b700bad1 gptoss optim fixes (#17356) 2026-08-01 09:10:06 -07:00
chenyuandGitHub 8e524ca467 CONST related cleanups [pr] (#17352)
ConstFloat(nan) != nan should be False, and some Invalid bool cleanups
2026-08-01 09:54:28 -04:00
nimlgenandGitHub 665822ab34 hcq2: faster replace (#17353)
* hcq2: use runtime device for submit params

* hcq2: parameterize buffers in linear time
2026-08-01 16:32:04 +03:00
chenyuandGitHub 6c0ec39279 UOp.is_invalid [PR] (#17351)
helper to prep ConstArg
2026-08-01 02:36:32 -04:00
chenyuandGitHub b502fc1367 more const arg -> val (#17350) 2026-08-01 02:18:10 -04:00
qazalandGitHub 161783d8f7 add _device_num back to ast.variables (kimi) (#17327) 2026-08-01 14:56:50 +09:00
George HotzandGitHub 5a1c641f79 more arg -> val (#17349)
* more arg -> val

* kimi

* more
2026-07-31 22:44:24 -07:00
Adeeb ShihadehandGitHub 20b8ecff50 support new USB vendor ID (#17348) 2026-07-31 21:45:06 -07:00
George HotzandGitHub 9082ecef5d use .val to access the value of Ops.CONST (#17347) 2026-07-31 19:51:42 -07:00
George HotzandGitHub 099d69ff7d ci: split macos unit test into metal and mock runners (#17346) 2026-07-31 19:41:18 -07:00
George HotzandGitHub a88f832f0c remove UOp.val (#17345) 2026-07-31 18:58:38 -07:00
chenyuandGitHub 850989115d __int__ and __float__ work for weak (#17342) 2026-07-31 19:56:37 -04:00
sirhcmandGitHub 15d515299e heuristics: try multiple TC axes (#17341) 2026-07-31 19:25:18 -04:00
sirhcmandGitHub 85ced44db6 tc: don't allow reduce over output dims (#17340) 2026-07-31 16:48:51 -04:00
chenyuandGitHub 277433259e fix sym_infer for CAST (#17338) 2026-07-31 14:38:38 -04:00
chenyuandGitHub 8dc225e28e dtype_from_uop(INS) is None [PR] (#17337) 2026-07-31 14:33:15 -04:00
chenyuandGitHub ad24750487 const(value, dtype) -> const(value).cast(dtype) in tests (#17335) 2026-07-31 13:24:31 -04:00