-
Notifications
You must be signed in to change notification settings - Fork 3k
Insights: PaddlePaddle/PaddleNLP
Overview
Could not load contribution data
Please try again later
13 Pull requests merged by 8 people
-
[llm] support tensorwise fp8/int8 training
#10612 merged
Jun 5, 2025 -
improve qwen perf
#10692 merged
Jun 5, 2025 -
[Template] apply chat template support openai format.
#10701 merged
Jun 4, 2025 -
Update aistudio-sdk requirement
#10698 merged
Jun 4, 2025 -
Disable fast math due to precision issue
#10697 merged
Jun 4, 2025 -
limit stack use to prevent CUDA error 2
#10696 merged
Jun 4, 2025 -
add jinja plugin
#10677 merged
Jun 3, 2025 -
Act dequant possible Overflow fix
#10691 merged
Jun 3, 2025 -
Add fused_transpose_split_quant kernel
#10657 merged
Jun 3, 2025 -
Revert "[Auto-Parallel] optimize llama27b to avoid unnecessary communication"
#10682 merged
Jun 3, 2025 -
Update loading numpy format state dict
#10679 merged
May 30, 2025 -
[Auto-Parallel] Add benchmark for sharding opt in qwen N4C32 dy_auto
#10673 merged
May 30, 2025 -
[Auto-Parallel] optimize llama27b to avoid unnecessary communication
#10671 merged
May 30, 2025
20 Pull requests opened by 14 people
-
'nola提交'
#10676 opened
May 29, 2025 -
aistudio 使用 SDK 进行下载
#10678 opened
May 30, 2025 -
[feat] Integrate Galvatron (an automatic parallel system integrating …
#10680 opened
May 30, 2025 -
Revert "[Auto-Parallel] optimize llama27b to avoid unnecessary communication"
#10681 opened
May 30, 2025 -
Implement wintx_unzip as custom_op
#10683 opened
May 31, 2025 -
【Hackathon 8th No.30】在 PaddleNLP 中复现 Gemma2 模型
#10684 opened
May 31, 2025 -
【Hackathon 8th No.29】在 PaddleNLP 中复现 ModernBERT 模型
#10686 opened
Jun 1, 2025 -
【Hackathon 8th No.28】在 PaddleNLP 中复现 Phi3
#10688 opened
Jun 1, 2025 -
Support sm90 for wint4
#10689 opened
Jun 3, 2025 -
[MergeKit] update mergekit with remove_keys input for lora merge
#10693 opened
Jun 3, 2025 -
Tmp llama13b benchmark
#10694 opened
Jun 3, 2025 -
Llama auto 13b benchmark with fuse linear
#10695 opened
Jun 3, 2025 -
add register for auto model
#10699 opened
Jun 4, 2025 -
[Infra] add return_full_hidden_state
#10700 opened
Jun 4, 2025 -
【Inference Optimize】update moe_preprocess for wint2.x
#10702 opened
Jun 4, 2025 -
【Hackathon 8th No.31】在 PaddleNLP 中复现 Apollo 精调算法
#10703 opened
Jun 4, 2025 -
【Hackathon 8th No.32】在 PaddleNLP 中复现 Adam-mini 精调算法
#10704 opened
Jun 4, 2025 -
add generate_expert_indices op
#10705 opened
Jun 4, 2025 -
Support wint2 unzip
#10706 opened
Jun 5, 2025 -
Update llama_dygraph_auto_bs8_fp32_DP2 mem_base
#10707 opened
Jun 5, 2025
9 Issues closed by 1 person
-
[Question]: 我这边想部署个离线的微型nlp模型到我的exe中
#10238 closed
Jun 5, 2025 -
[Question]:多个预训练模型无法下载
#10042 closed
Jun 4, 2025 -
[Bug]: taskflow uie 动转静报错
#10159 closed
Jun 4, 2025 -
[Bug]: paddle_ops cpu 编译报错
#10199 closed
Jun 3, 2025 -
[Question]: LogitsProcessorList 缺少__iter__和extend方法
#9926 closed
Jun 2, 2025 -
[Question]: ernie-3-tiny 按文档步骤运行例子报错
#9896 closed
Jun 1, 2025 -
[Question]: doccano=1.6.2，autolabeling无法实现，一直转圈
#9744 closed
May 31, 2025 -
[Question]: Error loading layoutlmv2-base-uncased: Missing model_state.pdparams file
#9868 closed
May 31, 2025 -
[Bug]: UIE信息抽取预测时，请求多次，每次返回结果不固定
#10150 closed
May 31, 2025
1 Issue opened by 1 person
38 Unresolved conversations
Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.
-
Add llama-13b dynamic auto benchmark
#10654 commented on
Jun 5, 2025 • 1 new comment -
Dsk ep
#10311 commented on
May 31, 2025 • 0 new comments -
Group Gemm 适配 DeepEP low_latency
#10314 commented on
May 31, 2025 • 0 new comments -
Using single fp8 gemm
#10318 commented on
May 31, 2025 • 0 new comments -
Temporarily added Ep8 support for tokens_unzip_stable and tokens_zip op
#10326 commented on
Jun 1, 2025 • 0 new comments -
Optimize expert performance
#10328 commented on
Jun 2, 2025 • 0 new comments -
force select self node
#10331 commented on
Jun 2, 2025 • 0 new comments -
update doc
#10333 commented on
Jun 2, 2025 • 0 new comments -
[LLM INFER] refactor return_full_hidden_states
#10335 commented on
Jun 2, 2025 • 0 new comments -
Add fused expert
#10345 commented on
Jun 3, 2025 • 0 new comments -
【Hackathon 8th No.32】 Adam-mini 精调算法复现
#10413 commented on
Jun 4, 2025 • 0 new comments -
[Inference] Add new wint2.75/wint2.5 quant type and support DeepseekV3
#10578 commented on
Jun 3, 2025 • 0 new comments -
Integrate DataProto into the GRPO
#10597 commented on
May 30, 2025 • 0 new comments -
[CI] skip unit case for hang
#10642 commented on
Jun 4, 2025 • 0 new comments -
[workflow] Add workflow yaml
#10643 commented on
Jun 5, 2025 • 0 new comments -
llama_with_auto_pp
#10648 commented on
Jun 3, 2025 • 0 new comments -
Support fastsafetensors to load model
#10667 commented on
Jun 3, 2025 • 0 new comments -
refine fp8 code
#10669 commented on
Jun 5, 2025 • 0 new comments -
Wintx unzip
#10674 commented on
May 30, 2025 • 0 new comments -
[Question]:MTP原理
#10258 commented on
May 30, 2025 • 0 new comments -
完成实测、提交测评，赢取奖金！——DeepSeek-R1-MTP 单机部署实战
#10166 commented on
Jun 3, 2025 • 0 new comments -
[Question]: PP-UIE-7B在微调时为什么突然显存爆增，导致OOM，差不多80G显存都爆炸了
#10572 commented on
Jun 3, 2025 • 0 new comments -
[Bug]: 在用Taskflow推理的时候，指定本地模型路径没生效
#10660 commented on
Jun 5, 2025 • 0 new comments -
[Question]: Will open dataset for ppuie model
#10556 commented on
Jun 5, 2025 • 0 new comments -
fix(sec): upgrade aiohttp to 3.8.5
#7264 commented on
Jun 4, 2025 • 0 new comments -
Mix
#7547 commented on
Jun 2, 2025 • 0 new comments -
[Trainer] Add metrics dumper in background
#9112 commented on
Jun 5, 2025 • 0 new comments -
[Auto Parallel] fix loss sum for auto parallel
#9234 commented on
May 31, 2025 • 0 new comments -
LoriKiT
#9776 commented on
Jun 1, 2025 • 0 new comments -
longlora-paddle
#9939 commented on
Jun 3, 2025 • 0 new comments -
[Config] Add model configs
#9981 commented on
Jun 1, 2025 • 0 new comments -
[custom device]: replace broadcast kernel with custom api for sdaa.
#10082 commented on
Jun 5, 2025 • 0 new comments -
update_llama_conf_cinn_0319
#10198 commented on
Jun 4, 2025 • 0 new comments -
【Inference Optimize】Paddle supports MOE model EP parallel
#10201 commented on
Jun 3, 2025 • 0 new comments -
Integrate DeepGEMM into fused_moe op
#10210 commented on
Jun 3, 2025 • 0 new comments -
[WIP] Add deepep timer.
#10232 commented on
May 31, 2025 • 0 new comments -
【PPO】support dataproto & fix dataflow
#10259 commented on
Jun 1, 2025 • 0 new comments -
Dsv3 dev
#10273 commented on
Jun 4, 2025 • 0 new comments