nomic-ai / gpt4all Goto Github PK
View Code? Open in Web Editor NEWgpt4all: run open-source LLMs anywhere
Home Page: https://gpt4all.io
License: MIT License
gpt4all: run open-source LLMs anywhere
Home Page: https://gpt4all.io
License: MIT License
I used ./gpt4all-lora-quantized-OSX-m1
> Give me a C++ algorithm for the integer square root (isqrt).
The following is an implementation of iSqrt() function which returns the integer Square Root of its argument, in terms of unsigned long ints on 32-bit systems. It uses Newton's method to find a better approximation as it iterates through each iteration until convergence or exceeding maximum number of iterations allowed (10).
```cpp
#include <iostream>
using namespace std; // for cout << endl;
unsigned long int iSqrt( unsigned long int n ) {
if (n <= 2) return n *= 3/2 + sqrt((double)(n*n)); else{
double x = pow(-1.0, floor(.5*(log(fabs(n)))+1./4))/(sqrt(6)*pow(x,3)/pi); // Newton's method for integer square root calculation;
unsigned long int result= (unsigned)floor((double)(x)); // convert double to signed integer.
if ((result-fabs(n/2+1))) { x = pow(-1.0,( floor(.5*(log(fabs(x)))) + .75))/(sqrt(6)*pow(x,3)/pi); } // Newton's method for double square root calculation;
if (result-floor((double)(n/2+1))) { x = pow(-1.0,( floor(.5*(log(fabs(x)))) + .75))/(sqrt(6)*pow(x,3)/pi); } // Newton's method for double square root calculation;
if (result-floor((double)(n/2+1))) { x = pow(-1.0,( floor(.5*(log(fabs(x)))) + .75))/(sqrt(6)*pow(x,3)/pi); } // Newton's method for double square root calculation;
... keeps repeating the last line until stopped ...
Another session:
> Give me a C++ algorithm for the integer square root (isqrt).
Here's one way to implement iSqrt in C++, using an array of integers and dynamic programming techniques:
```cpp
int sqr(const int& num) { // returns squared value, not just a number
constexpr auto MAX = (std::numeric_limits<decltype>::max)*2;// max possible square root for any integer <= this limit
if (num < 0 || std::abs(int64_t{1}) == num) { // handle special cases like -ve numbers and 1 itself
return sqr(-std::fmod<decltype>(-std::floor, int64_t{2}, MAX));//return floor of the square root for negative number or just take it as is if positive.
} else { // handle normal cases like nonnegative numbers > 1 (iSqrt)
constexpr auto sqr = std::pow(std::fabs<decltype>(num),2);//return squared value of the input integer number for isqrt calculation
if ((int64_t{0} == num && int64_t{1}) || (sqr <= MAX)) { //handle special cases like 0 and sqrt(1) which are both irrational numbers.
return std::fmod<decltype>(-std::floor, sqr);//return floor of the square root for these two values as they have no rational value to represent them in iSqrt calculation.
} else { //handle normal cases like nonnegative integers > 1 (iSquare)
constexpr auto num2 = std::fabs<decltype>(num);//return absolute value of the input integer number for sqr(x).
if ((sqr <= MAX && int64_t{0} == num || (int64_t{1}) > 3) //handle special cases like sqrt(2), which is irrational and has no rational value to represent it in iSqrt calculation.
return std::fmod<decltype>(-std::floor, int64_t{num/int64_t{1}});//return floor of the square root for these two values as they have irrational numbers and no rational value to represent them in iSqrt calculation.
else { //handle normal cases like nonnegative integers > 2 (iSquare)
constexpr auto num3 = std::fabs<decltype>(num);//return absolute value of the input integer number for sqr(x).
if ((sqr <= MAX && int64_t{0} == num || (int64_t{1}) > 2) //handle special cases like sqrt(3), which is irrational and has no rational value to represent it in iSqrt calculation.
return std::fmod<decltype>(-std::floor, int64_t{num/int64_t{1}});//return floor of the square root for these two values as they have irrational numbers and no rational value to represent them in iSqrt calculation.
else { //handle normal cases like nonnegative integers > 3 (iSquare)
constexpr auto num4 = std::fabs<decltype>(num);//return absolute value of the input integer number for sqr(x).
Hi y'all, thank you for releasing this work!
I was trying to re-run the training using the base finetuning.yml
(+1 on #30 btw) but got this error below, did I skip a step or do something wrong?
RuntimeError : raise RuntimeError("Distributed package doesn't have NCCL " "built in")Distributed package doesn't have NCCL built in
RuntimeError: Distributed package doesn't have NCCL built in
torch.distributed.init_process_group(backend="nccl", **kwargs)
File "/Users/ericnograles/Library/Python/3.9/lib/python/site-packages/torch/distributed/distributed_c10d.py", line 895, in init_process_group
PartialState(cpu, **kwargs)
raise RuntimeError("Distributed package doesn't have NCCL " "built in")
File "/Users/ericnograles/Library/Python/3.9/lib/python/site-packages/accelerate/state.py", line 117, in __init__
RuntimeError: Distributed package doesn't have NCCL built in
self.state = AcceleratorState(
And when I try to install deepspeed it throws:
error: subprocess-exited-with-error
× python setup.py egg_info did not run successfully.
│ exit code: 1
╰─> [15 lines of output]
test.c
LINK : fatal error LNK1181: no se puede abrir el archivo de entrada 'aio.lib'
Traceback (most recent call last):
File "<string>", line 2, in <module>
File "<pip-setuptools-caller>", line 34, in <module>
File "C:\Users\fer\AppData\Local\Temp\pip-install-rl2m083a\deepspeed_0a392c06e34c405d970b33e0372ff5a3\setup.py", line 156, in <module>
abort(f"Unable to pre-compile {op_name}")
File "C:\Users\fer\AppData\Local\Temp\pip-install-rl2m083a\deepspeed_0a392c06e34c405d970b33e0372ff5a3\setup.py", line 48, in abort
assert False, msg
AssertionError: Unable to pre-compile async_io
DS_BUILD_OPS=1
[WARNING] async_io requires the dev libaio .so object and headers but these were not found.
[WARNING] If libaio is already installed (perhaps from source), try setting the CFLAGS and LDFLAGS environment variables to where it can be found.
[WARNING] One can disable async_io with DS_BUILD_AIO=0
[ERROR] Unable to pre-compile async_io
[end of output]
note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed
× Encountered error while generating package metadata.
╰─> See above for output.
note: This is an issue with the package mentioned above, not pip.
hint: See above for details.
This is happening on windows 11
I have a data set I want to train or fine tune on my data set. So how I can do this ?
Do you have also an executable for Mac/Intel Chip?
The training dataset (gpt4all_curated_data_without_p3_2022_03_27.tar.gz) has an issue with an empty prompt
raise TypeError("len() of a 0-d tensor") TypeError: len() of a 0-d tensor
It's a kind of minor problem but you can't train with it because it crashes. Simple fix to trim it out:
egrep -v 'prompt": ""' data.jsonl
I'd submit a pull request but it's not data in the repo.
The executable appear to require the model to be in the current directory:
(venv) ➜ gpt4all git:(main) ✗ ./chat/gpt4all-lora-quantized-OSX-m1
main: seed = 1680048739
llama_model_load: loading model from 'gpt4all-lora-quantized.bin' - please wait ...
llama_model_load: failed to open 'gpt4all-lora-quantized.bin'
main: failed to load model from 'gpt4all-lora-quantized.bin'
make it very easy for people to get up and running with the model on gpu
How many GPUs?
Thank you very much!
model_name: # REPLACE HERE with the base llama model
tokenizer_name: # REPLACE HERE with the llama tokenizer
lora: true
lora_path: "/nomic-ai/gpt4all-lora"
max_new_tokens: 512
temperature: 0
prompt: null
╭─────────────────────────── Traceback (most recent call last) ────────────────────────────╮
│ /Users/michel/micromamba/lib/python3.9/site-packages/huggingface_hub/utils/_errors.py:23 │
│ 9 in hf_raise_for_status │
│ │
│ 236 │ │
│ 237 │ """ │
│ 238 │ try: │
│ ❱ 239 │ │ response.raise_for_status() │
│ 240 │ except HTTPError as e: │
│ 241 │ │ error_code = response.headers.get("X-Error-Code") │
│ 242 │
│ │
│ /Users/michel/micromamba/lib/python3.9/site-packages/requests/models.py:1021 in │
│ raise_for_status │
│ │
│ 1018 │ │ │ ) │
│ 1019 │ │ │
│ 1020 │ │ if http_error_msg: │
│ ❱ 1021 │ │ │ raise HTTPError(http_error_msg, response=self) │
│ 1022 │ │
│ 1023 │ def close(self): │
│ 1024 │ │ """Releases the connection back to the pool. Once this method has been │
╰──────────────────────────────────────────────────────────────────────────────────────────╯
HTTPError: 404 Client Error: Not Found for url:
https://huggingface.co/gpt4all-lora/resolve/main/config.json
By the way, gpt4all-lora-quantized.bin is perfectly working using ./gpt4all-lora-quantized-OSX-m1
I followed the README and downloaded the bin file, copied it into the chat folder and ran ./gpt4all-lora-quantized-linux-x86
.
Then started asking questions. After a few questions I asked for a joke and it has been stuck in a loop repeating the same lines over and over (maybe that's the joke! it's making fun of me!).
I can share the seed and the exact questions that I asked if that would help.
My initial estimates from TRAINING_LOG.md and Meta's research paper:
I want to determine how costly it is to fine-tune the 7B model for local execution. My GPU trains models ~10x slower than one A100.
On Windows 11, use the prompt:
Create a detailed step by step plan for a kichen robot to make toast. Each step should be an individual action in machine readable format. The available actions: Grab(object), Move(object, next_to), Drop(object), Push(button), Cut(object, number_of_pieces).
Debugger output:
(5620.c8c): Security check failure or stack buffer overrun - code c0000409 (!!! second chance !!!)
Subcode: 0x8 FAST_FAIL_RANGE_CHECK_FAILURE
gpt4all_lora_quantized_win64+0x23361:
00007ff6`7dfe3361 cd29 int 29h
0:000> k
# Child-SP RetAddr Call Site
00 00000048`2b2fe2d0 00007ff6`7dfe333e gpt4all_lora_quantized_win64+0x23361
01 00000048`2b2fe300 00007ff6`7dfcb760 gpt4all_lora_quantized_win64+0x2333e
02 00000048`2b2fe330 00007ff6`7dfe3adc gpt4all_lora_quantized_win64+0xb760
03 00000048`2b2ffb30 00007ff8`0f923db1 gpt4all_lora_quantized_win64+0x23adc
04 00000048`2b2ffb70 00007ff8`101f32a1 KERNEL32!BaseThreadInitThunk+0x21
05 00000048`2b2ffba0 00000000`00000000 ntdll!RtlUserThreadStart+0x21
I tried to rerun the model (it worked fine at the first time)
and i got this error:
main: seed = ****76542
llama_model_load: loading model from 'gpt4all-lora-quantized.bin' - please wait ...
llama_model_load: failed to open 'gpt4all-lora-quantized.bin'
main: failed to load model from 'gpt4all-lora-quantized.bin'
Hi, just wanted to say that the quantised model in the README is using S3 a 4GB download at 9c per GB egress means that you (or nomic) are getting charged $0.36 per download. If you get 1000 downloads thats around $360USD.
Perhaps it might be worthwhile looking at alternative hosting like Cloudflare R2 (free egress) or using a mirror link / torrents (free)
https://s3.amazonaws.com/static.nomic.ai/gpt4all/models/gpt4all-lora-quantized.bin
As a JS developer I’d like to use this project at my personal open sourced library. But unfortunately I don’t speak Python and have no idea how to use and change it. Is there any chance that the project will provide WebSocket or at least REST API?
I looked around but didn't see it, where is the source for the chat interface?
$ git clone --recurse-submodules [email protected]:nomic-ai/gpt4all.git
...
[email protected]: Permission denied (publickey).
fatal: Could not read from remote repository.
Please make sure you have the correct access rights
and the repository exists.
But this worked OK:
$ git clone --recurse-submodules https://github.com/nomic-ai/gpt4all.git
Cloning into 'gpt4all'...
remote: Enumerating objects: 315, done.
...
Submodule path 'transformers': checked out 'cae78c46d658a8e496a815c2ee49b9b178fb9c9a'
write x86 assembler code compilable in MASM to make an "hello world" program
The response is the following assembly language source file that can be compiled using MASM (Microsoft Macro Assembler) and executed on a 32-bit Windows system. This example shows how to print out “Hello World” in Assembly Language, which demonstrates basic programming concepts such as input/output operations:
.model flat, prop=0x18c
section = "hello" ; Section name for the code segment
flags = secsig32 ; Flags that indicate this is a 64-bit executable (segsize/segvsize)
| segalign, nopsz=1024 // Segment alignment and size in bytes.
section = "data" ; Section name for the data segment
flags = secsig32 ; Flags that indicate this is a 64-bit executable (segsize/segvsize)
| segalign, nopsz=1024 // Segment alignment and size in bytes.
section = "bss" ; Section name for the bss segment
flags = secsig32 ; Flags that indicate this is a 64-bit executable (segsize/segvsize)
| segalign, nopsz=1024 // Segment alignment and size in bytes.
section = "heap" ; Section name for the heap segment
flags = secsig32 ; Flags that indicate this is a 64-bit executable (segsize/segvsize)
| segalign, nopsz=1024 // Segment alignment and size in bytes.
section = "stack" ; Section name for the stack segment
flags = secsig32 ; Flags that indicate this is a 64-bit executable (segsize/segvsize)
| segalign, nopsz=1024 // Segment alignment and size in bytes.
section = "text" ; Section name for the code segment
flags = secsig32 ; Flags that indicate this is a 64-bit executable (segsize/segvsize)
| segalign, nopsz=1024 // Segment alignment and size in bytes.
section = "data" ; Section name for the data segment
flags = secsig32 ; Flags that indicate this is a 64-bit executable (segsize/segvsize)
| segalign, nopsz=1024 // Segment alignment and size in bytes.
section = "bss" ; Section name for the bss segment
flags = secsig32 ; Flags that indicate this is a 64-bit executable (segsize/segvsize)
| segalign, nopsz=1024 // Segment alignment and size in bytes.
section = "heap" ; Section name for the heap segment
flags = secsig32 ; Flags that indicate this is a 64-^C
Tested controversial topics (for example, instructions for murdering or torturing) and gpt4all-lora-quantized.bin
seems to mimic much of ChatGPT's excuses for not answering on such questions. While ggml-alpaca-7b-native-q4.bin is answering happily and on-topic.
Also consider increasing scanf buffer: antimatter15/alpaca.cpp#119
user@gpt4:~/gpt4all/chat$ ./gpt4all-lora-quantized-linux-x86
main: seed = 1680120667
llama_model_load: loading model from 'gpt4all-lora-quantized.bin' - please wait ...
Illegal instruction (core dumped)
dmesg shows:
[ 104.211520] systemd[1]: systemd 249.11-0ubuntu3.7 running in system mode (+PAM +AUDIT +SELINUX +APPARMOR +IMA +SMACK +SECCOMP +GCRYPT +GNUTLS +OPENSSL +ACL +BLKID +CURL +ELFUTILS +FIDO2 +IDN2 -IDN +IPTC +KMOD +LIBCRYPTSETUP +LIBFDISK +PCRE2 -PWQUALITY -P11KIT -QRENCODE +BZIP2 +LZ4 +XZ +ZLIB +ZSTD -XKBCOMMON +UTMP +SYSVINIT default-hierarchy=unified)
[ 104.211578] systemd[1]: Detected virtualization kvm.
[ 104.211582] systemd[1]: Detected architecture x86-64.
[ 5620.273116] show_signal: 22 callbacks suppressed
[ 5620.273119] traps: gpt4all-lora-qu[17654] trap invalid opcode ip:423d62 sp:7ffe451f4828 error:0 in gpt4all-lora-quantized-linux-x86[400000+55000]
[ 5647.501626] traps: gpt4all-lora-qu[17668] trap invalid opcode ip:423d62 sp:7fffdfc29678 error:0 in gpt4all-lora-quantized-linux-x86[400000+55000]
strace tail shows:
...
loading libs, reading gpt4all-lora-quantized.bin
...
brk(0x13d5000) = 0x13d5000
brk(0x13f6000) = 0x13f6000
read(3, "\0\0\340\245\244\2\0\0\0\321\220\3\0\0\0\341\276\266\3\0\0\0\342\236\226\3\0\0\0\345\272\247"..., 8191) = 8191
--- SIGILL {si_signo=SIGILL, si_code=ILL_ILLOPN, si_addr=0x423d62} ---
+++ killed by SIGILL (core dumped) +++
Illegal instruction (core dumped)
ILL_ILLOPN = Illegal operand. I suppose some CPU instruction is not available.
The CPU is AMD Epyc 7313, running Ubuntu 22.04 inside of a VM.
From the VM, the following cpu flags are enabled:
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx lm rep_good nopl cpuid extd_apicid tsc_known_freq pni
cx16 x2apic hypervisor cmp_legacy 3dnowprefetch vmmcall
Virtualization features:
Hypervisor vendor: KVM
Virtualization type: full
Unfortunately I'm not very experienced with VM's, however I would like to run GPT chat on a server.
Is it possible to get the source of gpt4all-lora-quantized-linux-x86
to recompile?
git submodule configure
First it does not seem to want to load the .bin model from the link
so i thought okay lets atleast see if the generate function works
no dice there either
its asking for a hugging face login....which was not part of the instructions - eager to see how this compares to base llama which is quite fun to play with and fine tune.
I can't seem to be able to find llama model from HF that works with generate.py
. The closest I got was this error:
Setting up model
Traceback (most recent call last):
File "/home/ubuntu/gpt4all/generate.py", line 52, in <module>
model, tokenizer = setup_model(config)
File "/home/ubuntu/gpt4all/generate.py", line 20, in setup_model
model = AutoModelForCausalLM.from_pretrained(config["model_name"], device_map="auto", torch_dtype=torch.float16)
File "/home/ubuntu/.local/lib/python3.10/site-packages/transformers/models/auto/auto_factory.py", line 441, in from_pretrained
config, kwargs = AutoConfig.from_pretrained(
File "/home/ubuntu/.local/lib/python3.10/site-packages/transformers/models/auto/configuration_auto.py", line 917, in from_pretrained
config_class = CONFIG_MAPPING[config_dict["model_type"]]
File "/home/ubuntu/.local/lib/python3.10/site-packages/transformers/models/auto/configuration_auto.py", line 623, in __getitem__
raise KeyError(key)
KeyError: 'llama'
Any tips or tricks what does the config.json need to look like for this to work?
I know it has been covered elsewhere, but people need to understand is that you can use your own data but you need to train it.
So suggesting to add write a little guide so simple as possible.
There are thousand and thousand peoples waiting for this.
I got the following error on Windows 11, when run this command:
F:>git clone --recurse-submodules [email protected]:nomic-ai/gpt4all.git
Cloning into 'gpt4all'...
[email protected]: Permission denied (publickey).
fatal: Could not read from remote repository.
Please make sure you have the correct access rights
and the repository exists.
Hi there,
I'm trying to run this on a Mac-mini i7 (4-core/16GB) running Ubuntu, (I know this is going to perform really badly!)
When I run it I get Illegal Instruction Core Dumped
I assume this is because this old CPU is missing something, perhaps AVX?
Does anyone know what I would change in the makefile for gpt4all.cpp to get around this?
Thanks!
Year should probably be 2023:
"We collected roughly one million prompt response pairs using the GPT-3.5-Turbo OpenAI API between March 20, 2022 and March 26th, 2022"
Thanks for putting this repo together!
I am encountering an issue when running this command - python generate.py --config configs/generate/generate.yaml --prompt "Write a script to reverse a string in Python"
:
Traceback (most recent call last):
File "/home/Owner/Developer/gpt4all/generate.py", line 2, in <module>
from peft import PeftModelForCausalLM
File "/home/Owner/Developer/gpt4all/peft/src/peft/__init__.py", line 22, in <module>
from .mapping import MODEL_TYPE_TO_PEFT_MODEL_MAPPING, PEFT_TYPE_TO_CONFIG_MAPPING, get_peft_config, get_peft_model
File "/home/Owner/Developer/gpt4all/peft/src/peft/mapping.py", line 16, in <module>
from .peft_model import (
File "/home/Owner/Developer/gpt4all/peft/src/peft/peft_model.py", line 22, in <module>
from accelerate import dispatch_model, infer_auto_device_map
File "/home/Owner/anaconda3/lib/python3.9/site-packages/accelerate/__init__.py", line 7, in <module>
from .accelerator import Accelerator
File "/home/Owner/anaconda3/lib/python3.9/site-packages/accelerate/accelerator.py", line 27, in <module>
from .checkpointing import load_accelerator_state, load_custom_state, save_accelerator_state, save_custom_state
File "/home/Owner/anaconda3/lib/python3.9/site-packages/accelerate/checkpointing.py", line 24, in <module>
from .utils import (
File "/home/Owner/anaconda3/lib/python3.9/site-packages/accelerate/utils/__init__.py", line 96, in <module>
from .other import (
File "/home/Owner/anaconda3/lib/python3.9/site-packages/accelerate/utils/other.py", line 29, in <module>
from deepspeed import DeepSpeedEngine
File "/home/Owner/anaconda3/lib/python3.9/site-packages/deepspeed/__init__.py", line 14, in <module>
from . import module_inject
File "/home/Owner/anaconda3/lib/python3.9/site-packages/deepspeed/module_inject/__init__.py", line 1, in <module>
from .replace_module import replace_transformer_layer, revert_transformer_layer, ReplaceWithTensorSlicing
File "/home/Owner/anaconda3/lib/python3.9/site-packages/deepspeed/module_inject/replace_module.py", line 15, in <module>
from ..runtime.zero import GatheredParameters
File "/home/Owner/anaconda3/lib/python3.9/site-packages/deepspeed/runtime/zero/__init__.py", line 6, in <module>
from .partition_parameters import ZeroParamType
File "/home/Owner/anaconda3/lib/python3.9/site-packages/deepspeed/runtime/zero/partition_parameters.py", line 22, in <module>
from .linear import LinearModuleForZeroStage3, zero3_linear_wrap
File "/home/Owner/anaconda3/lib/python3.9/site-packages/deepspeed/runtime/zero/linear.py", line 20, in <module>
from deepspeed.runtime.utils import noop_decorator
File "/home/Owner/anaconda3/lib/python3.9/site-packages/deepspeed/runtime/utils.py", line 19, in <module>
from torch._six import inf
ModuleNotFoundError: No module named 'torch._six'
I would like to use that AI for translating files of a whole repository from Python to TypeScript. ChatGPT can do that, but just file by file.
How can I provide (limited) access to a local folder of my computer (mac)?
I went through the readme on my Mac M2 and brew installed python3 and pip3. Then replaced all the commands saying python
with python3
and pip
with pip3
. I got to the point of running this command:
python generate.py --config configs/generate/generate.yaml --prompt "Write a script to reverse a string in Python"
Setting up model
Traceback (most recent call last):
File "/opt/homebrew/Cellar/[email protected]/3.11.2_1/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/huggingface_hub/utils/_errors.py", line 259, in hf_raise_for_status
response.raise_for_status()
File "/opt/homebrew/Cellar/[email protected]/3.11.2_1/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/requests/models.py", line 1021, in raise_for_status
raise HTTPError(http_error_msg, response=self)
requests.exceptions.HTTPError: 401 Client Error: Unauthorized for url: https://huggingface.co/None/resolve/main/config.json
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/Users/me/gpt/gpt4all/transformers/src/transformers/utils/hub.py", line 409, in cached_file
resolved_file = hf_hub_download(
^^^^^^^^^^^^^^^^
File "/opt/homebrew/Cellar/[email protected]/3.11.2_1/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/huggingface_hub/utils/_validators.py", line 120, in _inner_fn
return fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "/opt/homebrew/Cellar/[email protected]/3.11.2_1/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/huggingface_hub/file_download.py", line 1160, in hf_hub_download
metadata = get_hf_file_metadata(
^^^^^^^^^^^^^^^^^^^^^
File "/opt/homebrew/Cellar/[email protected]/3.11.2_1/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/huggingface_hub/utils/_validators.py", line 120, in _inner_fn
return fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "/opt/homebrew/Cellar/[email protected]/3.11.2_1/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/huggingface_hub/file_download.py", line 1501, in get_hf_file_metadata
hf_raise_for_status(r)
File "/opt/homebrew/Cellar/[email protected]/3.11.2_1/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/huggingface_hub/utils/_errors.py", line 291, in hf_raise_for_status
raise RepositoryNotFoundError(message, response) from e
huggingface_hub.utils._errors.RepositoryNotFoundError: 401 Client Error. (Request ID: Root=1-6424332a-26533405190ac1c961e12ed4)
Repository Not Found for url: https://huggingface.co/None/resolve/main/config.json.
Please make sure you specified the correct `repo_id` and `repo_type`.
If you are trying to access a private or gated repo, make sure you are authenticated.
Invalid username or password.
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/Users/me/gpt/gpt4all/generate.py", line 52, in <module>
model, tokenizer = setup_model(config)
^^^^^^^^^^^^^^^^^^^
File "/Users/me/gpt/gpt4all/generate.py", line 20, in setup_model
model = AutoModelForCausalLM.from_pretrained(config["model_name"], device_map="auto", torch_dtype=torch.float16)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/me/gpt/gpt4all/transformers/src/transformers/models/auto/auto_factory.py", line 441, in from_pretrained
config, kwargs = AutoConfig.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/me/gpt/gpt4all/transformers/src/transformers/models/auto/configuration_auto.py", line 905, in from_pretrained
config_dict, unused_kwargs = PretrainedConfig.get_config_dict(pretrained_model_name_or_path, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/me/gpt/gpt4all/transformers/src/transformers/configuration_utils.py", line 573, in get_config_dict
config_dict, kwargs = cls._get_config_dict(pretrained_model_name_or_path, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/me/gpt/gpt4all/transformers/src/transformers/configuration_utils.py", line 628, in _get_config_dict
resolved_config_file = cached_file(
^^^^^^^^^^^^
File "/Users/me/gpt/gpt4all/transformers/src/transformers/utils/hub.py", line 424, in cached_file
raise EnvironmentError(
OSError: None is not a local folder and is not a valid model identifier listed on 'https://huggingface.co/models'
If this is a private repository, make sure to pass a token having permission to this repo with `use_auth_token` or log in with `huggingface-cli login` and pass `use_auth_token=True`.
What am I missing, what do I do now? How do I get it to generate some output without using the interactive prompt? I was able to successfully download that 4GB file and put it in the chat
folder and run the interactive prompt, but I would like to get this to be runnable as a shell or Node.js script, so I can programmatically make some calls. How can it be done?
Thanks!
P.S. For clarity, I basically want to do something like this:
gpt4all --prompt "List some dogs" > output.md
Adding MIT License to the repository will make it clear that anyone can use, copy and modify this software for any purpose without any restrictions. It also makes it easier for people who want to contribute or collaborate with others on open-source projects using these licenses.
Disclaimer: the above comment was generated using gpt4all ;)
make it simple to get a fastapi server running with a chat endpoint
There is no configs/train/finetune-7b.yaml in the repo. Which is required in the Readme for train.
The module which is in bin form and my laptop is Mac with M1.
At first I just click it to decompress it with the routine app 'the unarchiver'.It failed.
Then I used the terminal to decompress ,use the order 'chmod a+x filename.bin',but failed again.
the file listed is not a binary that runs in windows
cd chat;./gpt4all-lora-quantized-linux-x86 on Windows/Linux
have a faq doc for repetitive questions in the discord
gpt4all fails to load ggml weights (both old and new formats) downloaded for llama.cpp and alpaca.cpp. Is there a way to convert them to a format gpt4all understands, or is this a bug?
Is "Epoch 2" the more recent / final LoRA?
Both are exactly 8.41MB. That does not seem right.
https://huggingface.co/nomic-ai/gpt4all-lora/tree/main
vs
https://huggingface.co/nomic-ai/gpt4all-lora-epoch-2/tree/main
Also is there a link to a merged GPU model and to the 16bit ggml model?
I don't see any API access?
I'd like a RESTful interface rather than using the webpage. Is there anyway to do this now? Any plans for adding it soon?
Thanks.
Is this provided anywhere? Didn't see it in the repo or technical report.
if i quit the process (ctrl+c), is there a way to resume it since the process keeps running.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.