swift15: all experts through GPTQ w4 (plugin one-table-one-encoding law)

rad_arch_declare rejects the split tier (down block-fp8 + gate_up w4): the
qwen4exp_fp8 kernel serves one encoding per expert table (w4nl64a8h) with only
whole bf16 experts beside it. Recipe now sends gate_up AND down through the
GPTQ w4 rule (down calibrated by the Grams, gate_up uncalibrated), drops the
down-fp8 rule, and lowers the size floor 126000->108000 MiB for the ~19 GiB
lighter container. Convert script: checkpoint lives under models/ukisai/, and
the in-container path must keep the subdir (basename stripped it).
This commit is contained in:
Frank committed 2026-10-07 21:31:10 -07:00
1 parent 2137080d1c
commit 73c37f4301
2 files changed
+25 -27

No files matched your search

+5 -4
View File
@@ -21,13 +21,14 @@ set -euo pipefail
PIN_1013=sha256:4c3f484564bb2872784e46b0a59cbd71becc9c1e746e301a3596ffae253c322e
IMAGE="stilldeadcode/radiance@$PIN_1013"
CKPT=/home/dank/models/Swift1.5-Qwen3.8-Flash-Next
CKPT=/home/dank/models/ukisai/Swift1.5-Qwen3.8-Flash-Next
RECIPE=/home/dank/serverAppConfigs-swift15/scripts/recipes/qwen4exp-swift15-w4dn8-i8-hc8m.recipe
CALIB=/home/dank/corpora/calib
OUT=/home/dank/models/rad/qwen4exp-swift15-w4dn8-i8-hc8m.rad
LOG=/home/dank/logs/swift15_convert.log
# gate_up ~37.6 + down ~37.7 + PLE 47.7 + trunk/embeddings/head ~10 GiB.
MIN_MIB=${FRANK_SWIFT15_RAD_MIN_MIB:-126000}
# gate_up ~37.6 + down ~18.8 + PLE 47.7 + trunk/embeddings/head ~10 GiB (all-w4 expert table;
# the plugin's one-encoding law dropped the block-fp8 down tier -- see the recipe header).
MIN_MIB=${FRANK_SWIFT15_RAD_MIN_MIB:-108000}
# Gate 1: inputs. A missing Gram directory converts UNCALIBRATED with a
# per-tensor stderr note, which is silently worse than failing here.
@@ -55,7 +56,7 @@ podman run --rm --name swift15-convert \
-v "$CALIB:/data/calib:ro" \
-e CALIB=/data/calib \
--entrypoint /opt/radiance/bin/rad-convert \
"$IMAGE" /models/"$(basename "$CKPT")" \
"$IMAGE" "/models/${CKPT#/home/dank/models/}" \
--recipe /campaign/"${RECIPE#/home/dank/serverAppConfigs-swift15/}" \
-o /models/rad/"$(basename "$OUT")" 2>&1 | tee -a "$LOG"
rc=${PIPESTATUS[0]}