# Latest **URL:** https://deeptalk.lambda.ai/latest.md?page=1 [Latest](https://deeptalk.lambda.ai/latest.md) · [Categories](https://deeptalk.lambda.ai/categories.md) **Page:** 2 --- ## [GH200 PyTorch installation failure](https://deeptalk.lambda.ai/t/gh200-pytorch-installation-failure/4521)
**Author:** [@isaac](https://deeptalk.lambda.ai/u/isaac)\ **Replies:** 1\ **Last updated:** [September 30, 2025, 1:04pm UTC](https://deeptalk.lambda.ai/t/gh200-pytorch-installation-failure/4521 "2025-09-30T13:04:42Z")
When I run the commands here, I get ERROR: No matching distribution found for torch. python -m venv vllm-env source vllm-env/bin/activate python -m pip install -U pip setuptools wheel pip install --pre torch --index-url… --- ## [Unable to create new instance](https://deeptalk.lambda.ai/t/unable-to-create-new-instance/4754)
**Author:** [@t.nguyen](https://deeptalk.lambda.ai/u/t.nguyen)\ **Replies:** 1\ **Last updated:** [September 22, 2025, 2:31am UTC](https://deeptalk.lambda.ai/t/unable-to-create-new-instance/4754 "2025-09-22T02:31:14Z")
When I click to Launch Instance button, the UI is frozen in this stage (see attached photo). Could you please take a look? Thanks! --- ## [Sunsetting Chat == Sunsetting Inference?](https://deeptalk.lambda.ai/t/sunsetting-chat-sunsetting-inference/4744)
**Author:** [@walter](https://deeptalk.lambda.ai/u/walter)\ **Replies:** 4\ **Last updated:** [September 16, 2025, 9:41am UTC](https://deeptalk.lambda.ai/t/sunsetting-chat-sunsetting-inference/4744 "2025-09-16T09:41:55Z")
“Lamba Chat has been sunsetted. Service will continue through September 25th.” \*\*Will Lambda Inference continue after September 25? \*\* --- ## [Session Persistence in Lambda Chat](https://deeptalk.lambda.ai/t/session-persistence-in-lambda-chat/4752)
**Author:** [@Anchained](https://deeptalk.lambda.ai/u/Anchained)\ **Replies:** 1\ **Last updated:** [September 15, 2025, 5:02am UTC](https://deeptalk.lambda.ai/t/session-persistence-in-lambda-chat/4752 "2025-09-15T05:02:56Z")
I was using Lambda Chat while in incognito mode and when I opened a new incognito tab or window my previous chats persisted even without signing in. How does this work under the hood? Can someone link a resource or give … --- ## [Booting takes long time and then "alert" status for gpu\_1x\_h100\_pcie](https://deeptalk.lambda.ai/t/booting-takes-long-time-and-then-alert-status-for-gpu-1x-h100-pcie/3735)
**Author:** [@znittzel](https://deeptalk.lambda.ai/u/znittzel)\ **Replies:** 4\ **Last updated:** [September 15, 2025, 5:00am UTC](https://deeptalk.lambda.ai/t/booting-takes-long-time-and-then-alert-status-for-gpu-1x-h100-pcie/3735 "2025-09-15T05:00:12Z")
Im trying to test out the “gpu\_1x\_h100\_pcie” but booting takes really long time until finally status change to “alert”. Is there something wrong with it? --- ## [PS only reaches 28v when ramp is applied](https://deeptalk.lambda.ai/t/ps-only-reaches-28v-when-ramp-is-applied/4750)
**Author:** [@dannyelloko20](https://deeptalk.lambda.ai/u/dannyelloko20)\ **Replies:** 0\ **Last updated:** [September 9, 2025, 1:47pm UTC](https://deeptalk.lambda.ai/t/ps-only-reaches-28v-when-ramp-is-applied/4750 "2025-09-09T13:47:58Z")
SW (TS and LV) sets PS to 28v but only is able to reach 26v when ramp is set to 10. I can play with the values and get to 28v but why? Load should read 0.2A (CL to 0.5A). When ramp is off, voltage levels drop to .03v an… --- ## [Where is the H100 PCI?](https://deeptalk.lambda.ai/t/where-is-the-h100-pci/4745)
**Author:** [@FrankZ](https://deeptalk.lambda.ai/u/FrankZ)\ **Replies:** 3\ **Last updated:** [September 5, 2025, 3:35pm UTC](https://deeptalk.lambda.ai/t/where-is-the-h100-pci/4745 "2025-09-05T15:35:50Z")
As title. I just used H100 PCI last night but today it seems this type of instance has been removed. --- ## [Using docker context with Lambda VM](https://deeptalk.lambda.ai/t/using-docker-context-with-lambda-vm/4748)
**Author:** [@BCF](https://deeptalk.lambda.ai/u/BCF)\ **Replies:** 1\ **Last updated:** [September 5, 2025, 10:08am UTC](https://deeptalk.lambda.ai/t/using-docker-context-with-lambda-vm/4748 "2025-09-05T10:08:19Z")
Hello, I’m looking to use the docker context feature with lambda cloud VMs. The problem seems to be that all docker commands in lambda must run with sudo and this probably blocks the docker context communication. I trie… --- ## [Feature Request: Spending Limits](https://deeptalk.lambda.ai/t/feature-request-spending-limits/4742)
**Author:** [@eeek](https://deeptalk.lambda.ai/u/eeek)\ **Replies:** 2\ **Last updated:** [September 1, 2025, 2:10pm UTC](https://deeptalk.lambda.ai/t/feature-request-spending-limits/4742 "2025-09-01T14:10:13Z")
Hello! I see a few topics where people asked if this functionality exists, but here is an official feature request. From the other topics: ”I’m afraid of leaving an instance accidentally running and racking up unexpect… --- ## [Ok ....I'll give a good THAT FIGURES!](https://deeptalk.lambda.ai/t/ok-ill-give-a-good-that-figures/4739)
**Author:** [@synacknet](https://deeptalk.lambda.ai/u/synacknet)\ **Replies:** 1\ **Last updated:** [September 1, 2025, 5:07am UTC](https://deeptalk.lambda.ai/t/ok-ill-give-a-good-that-figures/4739 "2025-09-01T05:07:39Z")
So I make an android app, basically a web wrapper for the lambda.chat I’m excited, when your on your android no more taking up your browser to use lambdas chat, create it, works well, looks good….then SLAP THEY ARE SHITT… --- ## [Are we being billed for data in the ".Trash" folder?](https://deeptalk.lambda.ai/t/are-we-being-billed-for-data-in-the-trash-folder/4737)
**Author:** [@banned4hax](https://deeptalk.lambda.ai/u/banned4hax)\ **Replies:** 3\ **Last updated:** [August 27, 2025, 5:43am UTC](https://deeptalk.lambda.ai/t/are-we-being-billed-for-data-in-the-trash-folder/4737 "2025-08-27T05:43:33Z")
I’ve been attempting to reduce the size of my persistent file drives to cut down on costs, but I’ve noticed that after deleting folders with 200+ GB worth of data, the reported file size has not changed as seen on https:… --- ## [Cheapest way to upload data into the filesystem with slow upload bandwidth](https://deeptalk.lambda.ai/t/cheapest-way-to-upload-data-into-the-filesystem-with-slow-upload-bandwidth/4473)
**Author:** [@dushan](https://deeptalk.lambda.ai/u/dushan)\ **Replies:** 3\ **Last updated:** [August 26, 2025, 7:30pm UTC](https://deeptalk.lambda.ai/t/cheapest-way-to-upload-data-into-the-filesystem-with-slow-upload-bandwidth/4473 "2025-08-26T19:30:52Z")
I just signed up and need to upload data into the lambdalabs filesystem before I can train on it. The docs state the only way to do this is by attaching the filesystem it to an instance. But the only instances available … --- ## [Is there a list of instance types by region somewhere?](https://deeptalk.lambda.ai/t/is-there-a-list-of-instance-types-by-region-somewhere/4722)
**Author:** [@utanapishtirox](https://deeptalk.lambda.ai/u/utanapishtirox)\ **Replies:** 5\ **Last updated:** [August 26, 2025, 7:28pm UTC](https://deeptalk.lambda.ai/t/is-there-a-list-of-instance-types-by-region-somewhere/4722 "2025-08-26T19:28:49Z")
I’m doing dev on an A100 in Virginia, but I want to be able to switch to an H100 or 8xH100, or even a cluster of B200s eventually. I’m trying to figure out the best region for locating my Filesystem instance, because it… --- ## [Inference API Privacy](https://deeptalk.lambda.ai/t/inference-api-privacy/4553)
**Author:** [@panopticon](https://deeptalk.lambda.ai/u/panopticon)\ **Replies:** 7\ **Last updated:** [August 22, 2025, 8:25pm UTC](https://deeptalk.lambda.ai/t/inference-api-privacy/4553 "2025-08-22T20:25:11Z")
What is the privacy guarantee of the Inference API? Are request/response content logged or stored in any way? --- ## [No display on Lambda Vector Pro](https://deeptalk.lambda.ai/t/no-display-on-lambda-vector-pro/4671)
**Author:** [@tandrew98](https://deeptalk.lambda.ai/u/tandrew98)\ **Replies:** 2\ **Last updated:** [August 20, 2025, 1:28am UTC](https://deeptalk.lambda.ai/t/no-display-on-lambda-vector-pro/4671 "2025-08-20T01:28:24Z")
I have a problem with the display on my machine. The BIOS (or anything) screen never comes up. It is just blank. Whichever input I select (DisplayPort or HDMI) my monitor says it does not find any input. The monitor and… --- ## [2 identical Lambda Quads; one can't connect to display; one can](https://deeptalk.lambda.ai/t/2-identical-lambda-quads-one-cant-connect-to-display-one-can/4669)
**Author:** [@bruce](https://deeptalk.lambda.ai/u/bruce)\ **Replies:** 1\ **Last updated:** [August 20, 2025, 1:26am UTC](https://deeptalk.lambda.ai/t/2-identical-lambda-quads-one-cant-connect-to-display-one-can/4669 "2025-08-20T01:26:56Z")
Hi. I have 2 identical Quads but only 1 can connect to my monitor. I need to be able to access the BIOS but am a bit stuck without the display. It’s an old VGA monitor and I get video signal via the 2nd HDMI port in the … --- ## [Are "Public Cloud" instances resources shared?](https://deeptalk.lambda.ai/t/are-public-cloud-instances-resources-shared/4695)
**Author:** [@tox46](https://deeptalk.lambda.ai/u/tox46)\ **Replies:** 1\ **Last updated:** [August 20, 2025, 1:22am UTC](https://deeptalk.lambda.ai/t/are-public-cloud-instances-resources-shared/4695 "2025-08-20T01:22:44Z")
When i create and launch an instance is the hardware share? and more specifically, are GPUs shared too? or when i launch an instance i get vGPUs unless on the “Private Cloud”? --- ## [Lambda Vector One Ram Upgrade](https://deeptalk.lambda.ai/t/lambda-vector-one-ram-upgrade/4701)
**Author:** [@stevenmoss](https://deeptalk.lambda.ai/u/stevenmoss)\ **Replies:** 1\ **Last updated:** [August 20, 2025, 1:02am UTC](https://deeptalk.lambda.ai/t/lambda-vector-one-ram-upgrade/4701 "2025-08-20T01:02:17Z")
Hi Team, I purchased a Lambda Vector One with a Ryzen 7000 series CPU configured with 64 GB RAM, I’d like to upgrade the RAM on the device, but read that Ryzen 7000 series tends to require certain products and configura… --- ## [Unable to login with google](https://deeptalk.lambda.ai/t/unable-to-login-with-google/4703)
**Author:** [@airgpu](https://deeptalk.lambda.ai/u/airgpu)\ **Replies:** 1\ **Last updated:** [August 20, 2025, 12:58am UTC](https://deeptalk.lambda.ai/t/unable-to-login-with-google/4703 "2025-08-20T00:58:37Z")
Hi I have a issue in logging in. My account is \[Moderator: email address redacted for privacy\] And I was trying to login with google. I’m still being billed, so I would appreciate a quick response --- ## [How to speed up the launch of calculations?](https://deeptalk.lambda.ai/t/how-to-speed-up-the-launch-of-calculations/4731)
**Author:** [@kaz133](https://deeptalk.lambda.ai/u/kaz133)\ **Replies:** 1\ **Last updated:** [August 20, 2025, 12:48am UTC](https://deeptalk.lambda.ai/t/how-to-speed-up-the-launch-of-calculations/4731 "2025-08-20T00:48:03Z")
I have a task literally 30 times a day to calculate the hash I need extremely quickly. And in principle there are no problems with this, but I cannot pay for the rest of the server time, that it will not do anything. Th… --- ## [VLLM cluster using VLLM production stack](https://deeptalk.lambda.ai/t/vllm-cluster-using-vllm-production-stack/4728)
**Author:** [@morrow\_maki](https://deeptalk.lambda.ai/u/morrow_maki)\ **Replies:** 0\ **Last updated:** [August 18, 2025, 7:15am UTC](https://deeptalk.lambda.ai/t/vllm-cluster-using-vllm-production-stack/4728 "2025-08-18T07:15:20Z")
Hello community, I’m looking forward to renting multiple lambdalabs machines (with H100 gpus) and deploying them on a cluster using the vllm production stack. My goal is to be able to manually purchase more machines fr… --- ## [How to choose the right batch size when starting deep learning training?](https://deeptalk.lambda.ai/t/how-to-choose-the-right-batch-size-when-starting-deep-learning-training/4726)
**Author:** [@sopelos](https://deeptalk.lambda.ai/u/sopelos)\ **Replies:** 0\ **Last updated:** [August 18, 2025, 5:50am UTC](https://deeptalk.lambda.ai/t/how-to-choose-the-right-batch-size-when-starting-deep-learning-training/4726 "2025-08-18T05:50:00Z")
Hello I am just getting started with deep learning and one of the things that confuses me is how to pick the right batch size for training.:slightly\_smiling\_face: I have seen tutorials where people use very small batch… --- ## [Is it possible to persist the workspace configuration?](https://deeptalk.lambda.ai/t/is-it-possible-to-persist-the-workspace-configuration/4309)
**Author:** [@jakobsteinbauer](https://deeptalk.lambda.ai/u/jakobsteinbauer)\ **Replies:** 2\ **Last updated:** [August 13, 2025, 6:41pm UTC](https://deeptalk.lambda.ai/t/is-it-possible-to-persist-the-workspace-configuration/4309 "2025-08-13T18:41:56Z")
When launching Lambda Cloud instances, I also connect a storage to it. I’ve seen that the content of the storage persists after restarting an instance, but that’s not true for my workspace configuration (like conda envi… --- ## [Adding a host based adapter card to a scalar server](https://deeptalk.lambda.ai/t/adding-a-host-based-adapter-card-to-a-scalar-server/4712)
**Author:** [@JayK](https://deeptalk.lambda.ai/u/JayK)\ **Replies:** 0\ **Last updated:** [August 1, 2025, 7:30pm UTC](https://deeptalk.lambda.ai/t/adding-a-host-based-adapter-card-to-a-scalar-server/4712 "2025-08-01T19:30:37Z")
Hello. I have already talked to support but they cannot assist. We purchased a new scalar server , basically this machine - SYS-421GE-TNRT | 4U | SuperServer | Products | Supermicro I want to add a Fiber Channel host … --- ## [Training Microsoft's MatterGen model doesn't use GPUs](https://deeptalk.lambda.ai/t/training-microsofts-mattergen-model-doesnt-use-gpus/4681)
**Author:** [@wigging](https://deeptalk.lambda.ai/u/wigging)\ **Replies:** 2\ **Last updated:** [July 10, 2025, 11:51pm UTC](https://deeptalk.lambda.ai/t/training-microsofts-mattergen-model-doesnt-use-gpus/4681 "2025-07-10T23:51:05Z")
I’m trying to train Microsoft’s mattergen model using eight A100 GPUs on the Lambda Cloud. The instance type I used is gpu\_8x\_a100\_80gb\_sxm4 and region is us-east-1. The Python version on the instance is 3.10.12. Here ar… --- ## [Tool calling in Lambda Inference API](https://deeptalk.lambda.ai/t/tool-calling-in-lambda-inference-api/4448)
**Author:** [@dhruvmullick\_aa](https://deeptalk.lambda.ai/u/dhruvmullick_aa)\ **Replies:** 6\ **Last updated:** [June 24, 2025, 4:36pm UTC](https://deeptalk.lambda.ai/t/tool-calling-in-lambda-inference-api/4448 "2025-06-24T16:36:48Z")
While using the Public Inference API, I get a blank output when using tools field, but legitimate output when not using tools in the data. This is the payload I’m using: { "model": "llama3.1-8b-instruct", … --- ## [Content made public](https://deeptalk.lambda.ai/t/content-made-public/4667)
**Author:** [@Masira](https://deeptalk.lambda.ai/u/Masira)\ **Replies:** 0\ **Last updated:** [June 17, 2025, 12:45pm UTC](https://deeptalk.lambda.ai/t/content-made-public/4667 "2025-06-17T12:45:47Z")
Hello. From what I understand, the broad license grant (worldwide, royalty-free, sublicensable, perpetual) applies when content is made public on the site such as public or unlisted demos. Does this also apply when usi… --- ## [Feature Request](https://deeptalk.lambda.ai/t/feature-request/4663)
**Author:** [@banned4hax](https://deeptalk.lambda.ai/u/banned4hax)\ **Replies:** 0\ **Last updated:** [June 15, 2025, 4:01pm UTC](https://deeptalk.lambda.ai/t/feature-request/4663 "2025-06-15T16:01:49Z")
I want to be able to name my terminal tabs in the jupyterlab web app, as I often work with groups of them in a single browser session. --- ## [No direct startup script support?](https://deeptalk.lambda.ai/t/no-direct-startup-script-support/4353)
**Author:** [@banned4hax](https://deeptalk.lambda.ai/u/banned4hax)\ **Replies:** 2\ **Last updated:** [June 15, 2025, 3:59pm UTC](https://deeptalk.lambda.ai/t/no-direct-startup-script-support/4353 "2025-06-15T15:59:11Z")
From what I can see in the forums, this post mentions using ansible to automate starting up a process on a Lambda GPU instance - is this really the best option we have at the moment? --- ## [Make the premium models free again](https://deeptalk.lambda.ai/t/make-the-premium-models-free-again/4655)
**Author:** [@b9162a02-be93-4451](https://deeptalk.lambda.ai/u/b9162a02-be93-4451)\ **Replies:** 4\ **Last updated:** [June 14, 2025, 12:39pm UTC](https://deeptalk.lambda.ai/t/make-the-premium-models-free-again/4655 "2025-06-14T12:39:00Z")
20 messages per month? wtf is this. it doesnt even say how you pay or how much it costs. [Previous page](https://deeptalk.lambda.ai/latest.md) [Next page](https://deeptalk.lambda.ai/latest.md?page=2)