Jan Bernlöhr

jbernloehr's activity

• All
• Models
• Datasets
• Spaces
• Buckets
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

New activity in Qwen/Qwen-72B 10 days ago

[BUG] Qwen-72B remote code is incompatible with Transformers 5.x

#12 opened 10 days ago by

jbernloehr

New activity in apple/OpenELM-1_1B-Instruct 14 days ago

OpenELM remote code is incompatible with transformers 5.x: config load fails with TypeError: OpenELMConfig.__post_init__() got an unexpected keyword argument 'use_cache'

#8 opened 14 days ago by

jbernloehr

New activity in Sandroeth/cali-0.1B about 1 month ago

modeling_cali.py: bf16 forward crashes with "expected scalar type Float but found BFloat16" at torch.matmul(att, v) — float32 RoPE cache promotes q/k

1

#2 opened about 1 month ago by

jbernloehr

New activity in ai-sage/GFusion-10B-A1.8B-base about 1 month ago

IndexError in _extract_kv_cache on empty DynamicCache during first forward (prefill) with use_cache=True

2

#1 opened about 1 month ago by

jbernloehr

New activity in vivo-ai/BlueLM-7B-Chat-32K 5 months ago

Fix softmax_scale type for torch.compile compatibility

#5 opened 5 months ago by

jbernloehr

New activity in deepseek-ai/deepseek-moe-16b-base 6 months ago

Fix compatibility with transformers 5.0

1

#8 opened 6 months ago by

jbernloehr