Spaces:
Running
white Screen in Web version
Made a mini game but its only running on phone and not on PC. If i Enter the Link with a PC i only get a white page
Can someone help me?
⚛️## Exception *🧠 User Action: 🤖/data/user/0/chatpdf.pro/files/Welcome to PDF AI🤖.pdf🚀 Try PDF AI now! Quickly read and process any PDF with our powerful AI features.
👉 https://pdfai.page.link/s1the trustworthiness, indicating RLAIF-V is complementary
with other types of feedback.
RLAIF-V produces generalizable high-quality feed-
back. We train different models with the feedback collected
during the first iteration of training RLAIF-V 12B. Specif-
ically, we train LLaVA 1.5 7B [33], LLaVA 1.5 13B [33],
MiniCPM-V [46], MiniCPM-V 2 [46]. with direct prefer-
ence optimization and report the trustworthiness improve-
ment in Figure 4. We observe that data collected from Om-
niLMM (as both the instruction model and labeler model)
with RLAIF-V framework can effectively reduce the hal-
lucination of other MLLMs on different benchmarks. No-
tably, the improvement can be even more significant com-
pared with the OmniLMM which generates the candidate
responses. The results demonstrate that feedback from
RLAIF.../data/user/0/chatpdf.pro/files/Welcome to PDF AI🤖.pdf🚀 Try PDF AI now! Quickly read and process any PDF with our powerful AI features.
👉 https://pdfai.page.link/s1the trustworthiness, indicating RLAIF-V is complementary
with other types of feedback.
RLAIF-V produces generalizable high-quality feed-
back. We train different models with the feedback collected
during the first iteration of training RLAIF-V 12B. Specif-
ically, we train LLaVA 1.5 7B [33], LLaVA 1.5 13B [33],
MiniCPM-V [46], MiniCPM-V 2 [46]. with direct prefer-
ence optimization and report the trustworthiness improve-
ment in Figure 4. We observe that data collected from Om-
niLMM (as both the instruction model and labeler model)
with RLAIF-V framework can effectively reduce the hal-
lucination of other MLLMs on different benchmarks. No-
tably, the improvement can be even more significant com-
pared with the OmniLMM which generates the candidate
responses. The results demonstrate that feedback from
RLAIF...