Fix 404s (#7363)

Signed-off-by: Olatunji Ruwase <tjruwase@gmail.com>

blogs/deepnvme/06-2025/README.md

@@ -24,7 +24,7 @@ We used DeepNVMe to develop FastPersist and ZeRO-Inference to target I/O bottlen
 ## FastPersist: Faster Model Checkpoint Creation
 Although saving model checkpoints to persistent storage is critical in model training, it is also a major bottleneck due to the inefficiencies of existing approaches. We developed [FastPersist](https://arxiv.org/abs/2406.13768) to address the performance challenges of checkpointing. FastPersist makes checkpointing overheads negligible during training through three key techniques: (i) DeepNVMe, (ii) data parallelism, and (iii) overlapping I/O and computation.
 
-Our goal here is to demonstrate the impact of DeepNVMe in FastPersist using single-process micro-benchmarks (available [here](https://github.com/deepspeedai/DeepSpeedExamples/tree/master/deepnvme/fastpersist)) which serialize a model checkpoint state from HBM to local NVMe. We use the popular PyTorch `torch.save()` as the baseline in our experiments, and integrate FastPersist into `torch.save()` to simplify adoption and performance comparisons.
+Our goal here is to demonstrate the impact of DeepNVMe in FastPersist using single-process micro-benchmarks (available [here](https://github.com/deepspeedai/DeepSpeedExamples/tree/master/deepnvme/model_checkpoint)) which serialize a model checkpoint state from HBM to local NVMe. We use the popular PyTorch `torch.save()` as the baseline in our experiments, and integrate FastPersist into `torch.save()` to simplify adoption and performance comparisons.
 
 ### Faster Saving of PyTorch Models to local NVMe Storage
 We measure the throughput of serializing Phi-3-Mini checkpoint state from HBM to local NVMe storage. The results are summarized in the Figure below. We observe significantly faster checkpointing with FastPersist compared to the baseline. We see speedups of over 20X in the 8xGen5 NVMe settings. We also observe FastPersist scaling with increased NVMe bandwidth of 8xGen5 compared with 4xGen5.
@@ -47,7 +47,7 @@ We measure the throughput of serializing Phi-3-Mini checkpoint state from HBM to
 
 
 ### Scaling HF Transformer Generation with Faster NVMe SSDs
-ZeRO-Inference enhances HF Transformer inference with efficient model offloading to DRAM or NVMe. We previously [evaluated](https://github.com/deepspeedai/DeepSpeed/blob/master/blogs/deepspeed-gds/README.md#high-performance-offloading-via-nvme-scaling) LLAMA-3-70B generation performance with NVMe offloading on a single GPU and four Gen4 NVMes in an Azure [NC_A100_v4](https://learn.microsoft.com/en-us/azure/virtual-machines/sizes/gpu-accelerated/nca100v4-series?tabs=sizebasic) VM. We measured the generation speed for a prompt of 512 tokens, output of 32 tokens, and batch size 96. Since NVMe bandwidth was the main bottleneck, we repeat the experiments on Azure ND-H200-v5 offering Gen5 NVMes. The results summarized in the Figure below show that ZeRO-Inference uses the increased NVMe bandwidths to improve generation speeds. For example, with GDS, generation speed improves from 7 tokens/sec with four Gen4 NVMes to 17 tokens/sec with four Gen5 NVMes, and further to 26 tokens/sec with eight Gen5 NVMes. We observe similar improvements without GDS. These results show that ZeRO-Inference performance can be improved in cost-effective manner by increasing NVMe bandwidths.
+ZeRO-Inference enhances HF Transformer inference with efficient model offloading to DRAM or NVMe. We previously [evaluated](https://github.com/deepspeedai/DeepSpeed/blob/master/blogs/deepnvme/08-2024/README.md#high-performance-offloading-via-nvme-scaling) LLAMA-3-70B generation performance with NVMe offloading on a single GPU and four Gen4 NVMes in an Azure [NC_A100_v4](https://learn.microsoft.com/en-us/azure/virtual-machines/sizes/gpu-accelerated/nca100v4-series?tabs=sizebasic) VM. We measured the generation speed for a prompt of 512 tokens, output of 32 tokens, and batch size 96. Since NVMe bandwidth was the main bottleneck, we repeat the experiments on Azure ND-H200-v5 offering Gen5 NVMes. The results summarized in the Figure below show that ZeRO-Inference uses the increased NVMe bandwidths to improve generation speeds. For example, with GDS, generation speed improves from 7 tokens/sec with four Gen4 NVMes to 17 tokens/sec with four Gen5 NVMes, and further to 26 tokens/sec with eight Gen5 NVMes. We observe similar improvements without GDS. These results show that ZeRO-Inference performance can be improved in cost-effective manner by increasing NVMe bandwidths.
 
 <img src="./media/hf_zinf_llama_70b.png">
 <div align="center">