PhalaCloud/GLM-5.3-W4AFP8
Text Generation • 386B • Updated • 2.12k • 10
AI & ML interests
Confidential AI
Recent Activity
View all activity
Papers
Implement Kubernetes Pod-Level Remote Attestation for Confidential Workloads on dstack
LegalCiteBench: Evaluating Citation Reliability in Legal Language Models
Is it possible to build a GLM-5.3-Flash-W4AFP8?
1
#1 opened about 2 months ago
by 
iwoolk
The performance of glm
1
#2 opened about 2 months ago
by 
jianjungki
h4x3rotab
updated a
model 2 months ago
W4AFP8 for 5.3
6
#7 opened 3 months ago
by 
jlondon
h4x3rotab
published a
model 2 months ago
h4x3rotab
updated a
model 3 months ago
h4x3rotab
published a
model 3 months ago
h4x3rotab
updated a
model 3 months ago
ERROR: Should have a `model_type` key in its config.json
1
#5 opened 4 months ago
by 
PakJoeng
8xH20 141GB cuda out of memory
1
#6 opened 4 months ago
by 
ErisLU
shelvenzhou
authored 6
papers 4 months ago
LegalCiteBench: Evaluating Citation Reliability in Legal Language Models
Paper • 2605.10186 • Published • 1
PASS: A Provenanced Access Subaccount System for Blockchain Wallets
Paper • 2604.22602 • Published • 1
Implement Kubernetes Pod-Level Remote Attestation for Confidential Workloads on dstack
Paper • 2606.03323 • Published • 1
Dstack: A Zero Trust Framework for Confidential Containers
Paper • 2509.11555 • Published • 1
Confidential Computing on NVIDIA Hopper GPUs: A Performance Benchmark Study
Paper • 2409.03992 • Published • 1
Persistent BitTorrent Trackers
Paper • 2511.17260 • Published • 1
marvintong
updated a
Space 4 months ago
Any chances for A100?
1
#4 opened 4 months ago
by 
traphix
GLM-5.2-W4AFP8 on 8×H100: fp8_e4m3 KV cache produces corrupted output, while BF16 KV works correctly
1
#3 opened 4 months ago
by 
loveblairsky
Is the quantize script opensource?
3
#2 opened 4 months ago
by 
wxsm