first commit
This commit is contained in:
antimatter15
parent
9b4a15b17d
commit
e750735216
4 changed files with 102 additions and 249 deletions
4
Makefile
4
Makefile
|
|
@ -195,6 +195,10 @@ main: main.cpp ggml.o utils.o
|
||||||
$(CXX) $(CXXFLAGS) main.cpp ggml.o utils.o -o main $(LDFLAGS)
|
$(CXX) $(CXXFLAGS) main.cpp ggml.o utils.o -o main $(LDFLAGS)
|
||||||
./main -h
|
./main -h
|
||||||
|
|
||||||
|
chat: chat.cpp ggml.o utils.o
|
||||||
|
$(CXX) $(CXXFLAGS) chat.cpp ggml.o utils.o -o chat $(LDFLAGS)
|
||||||
|
|
||||||
|
|
||||||
quantize: quantize.cpp ggml.o utils.o
|
quantize: quantize.cpp ggml.o utils.o
|
||||||
$(CXX) $(CXXFLAGS) quantize.cpp ggml.o utils.o -o quantize $(LDFLAGS)
|
$(CXX) $(CXXFLAGS) quantize.cpp ggml.o utils.o -o quantize $(LDFLAGS)
|
||||||
|
|
||||||
|
|
|
||||||
235
README.md
235
README.md
|
|
@ -1,223 +1,46 @@
|
||||||
# llama.cpp
|
# Alpaca.cpp
|
||||||
|
|
||||||
[](https://github.com/ggerganov/llama.cpp/actions)
|
Run a fast ChatGPT-like model locally on your device. The screencast below is not sped up and running on an M2 Macbook Air with 4GB of weights.
|
||||||
[](https://opensource.org/licenses/MIT)
|
|
||||||
|
|
||||||
Inference of [LLaMA](https://arxiv.org/abs/2302.13971) model in pure C/C++
|
|
||||||
|
|
||||||
**Hot topics:**
|
|
||||||
|
|
||||||
- Cache input prompts for faster initialization: https://github.com/ggerganov/llama.cpp/issues/64
|
|
||||||
- Create a `llama.cpp` logo: https://github.com/ggerganov/llama.cpp/issues/105
|
|
||||||
|
|
||||||
## Description
|
|
||||||
|
|
||||||
The main goal is to run the model using 4-bit quantization on a MacBook
|
|
||||||
|
|
||||||
- Plain C/C++ implementation without dependencies
|
|
||||||
- Apple silicon first-class citizen - optimized via ARM NEON
|
|
||||||
- AVX2 support for x86 architectures
|
|
||||||
- Mixed F16 / F32 precision
|
|
||||||
- 4-bit quantization support
|
|
||||||
- Runs on the CPU
|
|
||||||
|
|
||||||
This was [hacked in an evening](https://github.com/ggerganov/llama.cpp/issues/33#issuecomment-1465108022) - I have no idea if it works correctly.
|
|
||||||
Please do not make conclusions about the models based on the results from this implementation.
|
|
||||||
For all I know, it can be completely wrong. This project is for educational purposes.
|
|
||||||
New features will probably be added mostly through community contributions.
|
|
||||||
|
|
||||||
Supported platforms:
|
|
||||||
|
|
||||||
- [X] Mac OS
|
|
||||||
- [X] Linux
|
|
||||||
- [X] Windows (via CMake)
|
|
||||||
|
|
||||||
---
|
|
||||||
|
|
||||||
Here is a typical run using LLaMA-7B:
|
|
||||||
|
|
||||||
```java
|
|
||||||
make -j && ./main -m ./models/7B/ggml-model-q4_0.bin -p "Building a website can be done in 10 simple steps:" -t 8 -n 512
|
|
||||||
I llama.cpp build info:
|
|
||||||
I UNAME_S: Darwin
|
|
||||||
I UNAME_P: arm
|
|
||||||
I UNAME_M: arm64
|
|
||||||
I CFLAGS: -I. -O3 -DNDEBUG -std=c11 -fPIC -pthread -DGGML_USE_ACCELERATE
|
|
||||||
I CXXFLAGS: -I. -I./examples -O3 -DNDEBUG -std=c++11 -fPIC -pthread
|
|
||||||
I LDFLAGS: -framework Accelerate
|
|
||||||
I CC: Apple clang version 14.0.0 (clang-1400.0.29.202)
|
|
||||||
I CXX: Apple clang version 14.0.0 (clang-1400.0.29.202)
|
|
||||||
|
|
||||||
make: Nothing to be done for `default'.
|
|
||||||
main: seed = 1678486056
|
|
||||||
llama_model_load: loading model from './models/7B/ggml-model-q4_0.bin' - please wait ...
|
|
||||||
llama_model_load: n_vocab = 32000
|
|
||||||
llama_model_load: n_ctx = 512
|
|
||||||
llama_model_load: n_embd = 4096
|
|
||||||
llama_model_load: n_mult = 256
|
|
||||||
llama_model_load: n_head = 32
|
|
||||||
llama_model_load: n_layer = 32
|
|
||||||
llama_model_load: n_rot = 128
|
|
||||||
llama_model_load: f16 = 2
|
|
||||||
llama_model_load: n_ff = 11008
|
|
||||||
llama_model_load: ggml ctx size = 4529.34 MB
|
|
||||||
llama_model_load: memory_size = 512.00 MB, n_mem = 16384
|
|
||||||
llama_model_load: .................................... done
|
|
||||||
llama_model_load: model size = 4017.27 MB / num tensors = 291
|
|
||||||
|
|
||||||
main: prompt: 'Building a website can be done in 10 simple steps:'
|
|
||||||
main: number of tokens in prompt = 15
|
|
||||||
1 -> ''
|
|
||||||
8893 -> 'Build'
|
|
||||||
292 -> 'ing'
|
|
||||||
263 -> ' a'
|
|
||||||
4700 -> ' website'
|
|
||||||
508 -> ' can'
|
|
||||||
367 -> ' be'
|
|
||||||
2309 -> ' done'
|
|
||||||
297 -> ' in'
|
|
||||||
29871 -> ' '
|
|
||||||
29896 -> '1'
|
|
||||||
29900 -> '0'
|
|
||||||
2560 -> ' simple'
|
|
||||||
6576 -> ' steps'
|
|
||||||
29901 -> ':'
|
|
||||||
|
|
||||||
sampling parameters: temp = 0.800000, top_k = 40, top_p = 0.950000
|
|
||||||
|
|
||||||
|
|
||||||
Building a website can be done in 10 simple steps:
|
[](https://asciinema.org/a/dfJ8QXZ4u978Ona59LPEldtKK)
|
||||||
1) Select a domain name and web hosting plan
|
|
||||||
2) Complete a sitemap
|
|
||||||
3) List your products
|
|
||||||
4) Write product descriptions
|
|
||||||
5) Create a user account
|
|
||||||
6) Build the template
|
|
||||||
7) Start building the website
|
|
||||||
8) Advertise the website
|
|
||||||
9) Provide email support
|
|
||||||
10) Submit the website to search engines
|
|
||||||
A website is a collection of web pages that are formatted with HTML. HTML is the code that defines what the website looks like and how it behaves.
|
|
||||||
The HTML code is formatted into a template or a format. Once this is done, it is displayed on the user's browser.
|
|
||||||
The web pages are stored in a web server. The web server is also called a host. When the website is accessed, it is retrieved from the server and displayed on the user's computer.
|
|
||||||
A website is known as a website when it is hosted. This means that it is displayed on a host. The host is usually a web server.
|
|
||||||
A website can be displayed on different browsers. The browsers are basically the software that renders the website on the user's screen.
|
|
||||||
A website can also be viewed on different devices such as desktops, tablets and smartphones.
|
|
||||||
Hence, to have a website displayed on a browser, the website must be hosted.
|
|
||||||
A domain name is an address of a website. It is the name of the website.
|
|
||||||
The website is known as a website when it is hosted. This means that it is displayed on a host. The host is usually a web server.
|
|
||||||
A website can be displayed on different browsers. The browsers are basically the software that renders the website on the user’s screen.
|
|
||||||
A website can also be viewed on different devices such as desktops, tablets and smartphones. Hence, to have a website displayed on a browser, the website must be hosted.
|
|
||||||
A domain name is an address of a website. It is the name of the website.
|
|
||||||
A website is an address of a website. It is a collection of web pages that are formatted with HTML. HTML is the code that defines what the website looks like and how it behaves.
|
|
||||||
The HTML code is formatted into a template or a format. Once this is done, it is displayed on the user’s browser.
|
|
||||||
A website is known as a website when it is hosted
|
|
||||||
|
|
||||||
main: mem per token = 14434244 bytes
|
|
||||||
main: load time = 1332.48 ms
|
|
||||||
main: sample time = 1081.40 ms
|
|
||||||
main: predict time = 31378.77 ms / 61.41 ms per token
|
|
||||||
main: total time = 34036.74 ms
|
|
||||||
```
|
|
||||||
|
|
||||||
And here is another demo of running both LLaMA-7B and [whisper.cpp](https://github.com/ggerganov/whisper.cpp) on a single M1 Pro MacBook:
|
This combines the [LLaMA foundation model](https://github.com/facebookresearch/llama) with an [open reproduction](https://github.com/tloen/alpaca-lora) of [Stanford Alpaca](https://github.com/tatsu-lab/stanford_alpaca) a fine-tuning of the base model to obey instructions (akin to the [RLHF](https://huggingface.co/blog/rlhf) used to train ChatGPT).
|
||||||
|
|
||||||
https://user-images.githubusercontent.com/1991296/224442907-7693d4be-acaa-4e01-8b4f-add84093ffff.mp4
|
## Get started
|
||||||
|
|
||||||
## Usage
|
|
||||||
|
|
||||||
Here are the step for the LLaMA-7B model:
|
|
||||||
|
|
||||||
```bash
|
|
||||||
# build this repo
|
|
||||||
git clone https://github.com/ggerganov/llama.cpp
|
|
||||||
cd llama.cpp
|
|
||||||
make
|
|
||||||
|
|
||||||
# obtain the original LLaMA model weights and place them in ./models
|
|
||||||
ls ./models
|
|
||||||
65B 30B 13B 7B tokenizer_checklist.chk tokenizer.model
|
|
||||||
|
|
||||||
# install Python dependencies
|
|
||||||
python3 -m pip install torch numpy sentencepiece
|
|
||||||
|
|
||||||
# convert the 7B model to ggml FP16 format
|
|
||||||
python3 convert-pth-to-ggml.py models/7B/ 1
|
|
||||||
|
|
||||||
# quantize the model to 4-bits
|
|
||||||
./quantize.sh 7B
|
|
||||||
|
|
||||||
# run the inference
|
|
||||||
./main -m ./models/7B/ggml-model-q4_0.bin -t 8 -n 128
|
|
||||||
```
|
|
||||||
|
|
||||||
When running the larger models, make sure you have enough disk space to store all the intermediate files.
|
|
||||||
|
|
||||||
TODO: add model disk/mem requirements
|
|
||||||
|
|
||||||
### Interactive mode
|
|
||||||
|
|
||||||
If you want a more ChatGPT-like experience, you can run in interactive mode by passing `-i` as a parameter.
|
|
||||||
In this mode, you can always interrupt generation by pressing Ctrl+C and enter one or more lines of text which will be converted into tokens and appended to the current context. You can also specify a *reverse prompt* with the parameter `-r "reverse prompt string"`. This will result in user input being prompted whenever the exact tokens of the reverse prompt string are encountered in the generation. A typical use is to use a prompt which makes LLaMa emulate a chat between multiple users, say Alice and Bob, and pass `-r "Alice:"`.
|
|
||||||
|
|
||||||
Here is an example few-shot interaction, invoked with the command
|
|
||||||
```
|
|
||||||
./main -m ./models/13B/ggml-model-q4_0.bin -t 8 -n 256 --repeat_penalty 1.0 --color -i -r "User:" \
|
|
||||||
-p \
|
|
||||||
"Transcript of a dialog, where the User interacts with an Assistant named Bob. Bob is helpful, kind, honest, good at writing, and never fails to answer the User's requests immediately and with precision.
|
|
||||||
|
|
||||||
User: Hello, Bob.
|
|
||||||
Bob: Hello. How may I help you today?
|
|
||||||
User: Please tell me the largest city in Europe.
|
|
||||||
Bob: Sure. The largest city in Europe is Moscow, the capital of Russia.
|
|
||||||
User:"
|
|
||||||
|
|
||||||
```
|
```
|
||||||
Note the use of `--color` to distinguish between user input and generated text.
|
git clone https://github.com/antimatter15/alpaca.cpp
|
||||||
|
cd alpaca.cpp
|
||||||
|
|
||||||
|
make chat
|
||||||
|
./chat
|
||||||
### Android
|
|
||||||
|
|
||||||
You can easily run `llama.cpp` on Android device with [termux](https://play.google.com/store/apps/details?id=com.termux).
|
|
||||||
First, obtain the [Android NDK](https://developer.android.com/ndk) and then build with CMake:
|
|
||||||
```
|
```
|
||||||
$ mkdir build-android
|
|
||||||
$ cd build-android
|
You can download the weights for `ggml-alpaca-7b-14.bin` with BitTorrent `magnet:?xt=urn:btih:5aaceaec63b03e51a98f04fd5c42320b2a033010&dn=ggml-alpaca-7b-q4.bin&tr=udp%3A%2F%2Ftracker.opentrackr.org%3A1337%2Fannounce&tr=udp%3A%2F%2Fopentracker.i2p.rocks%3A6969%2Fannounce`
|
||||||
$ export NDK=<your_ndk_directory>
|
|
||||||
$ cmake -DCMAKE_TOOLCHAIN_FILE=$NDK/build/cmake/android.toolchain.cmake -DANDROID_ABI=arm64-v8a -DANDROID_PLATFORM=android-23 -DCMAKE_C_FLAGS=-march=armv8.4a+dotprod ..
|
|
||||||
$ make
|
Alternatively you can download them with IPFS.
|
||||||
|
|
||||||
|
```
|
||||||
|
# any of these will work
|
||||||
|
wget -O ggml-alpaca-7b-q4.bin -c https://gateway.estuary.tech/gw/ipfs/QmQ1bf2BTnYxq73MFJWu1B7bQ2UD6qG7D7YDCxhTndVkPC
|
||||||
|
wget -O ggml-alpaca-7b-q4.bin -c https://ipfs.io/ipfs/QmQ1bf2BTnYxq73MFJWu1B7bQ2UD6qG7D7YDCxhTndVkPC
|
||||||
|
wget -O ggml-alpaca-7b-q4.bin -c https://cloudflare-ipfs.com/ipfs/QmQ1bf2BTnYxq73MFJWu1B7bQ2UD6qG7D7YDCxhTndVkPC
|
||||||
```
|
```
|
||||||
Install [termux](https://play.google.com/store/apps/details?id=com.termux) on your device and run `termux-setup-storage` to get access to your SD card.
|
|
||||||
Finally, copy the `llama` binary and the model files to your device storage. Here is a demo of an interactive session running on Pixel 5 phone:
|
|
||||||
|
|
||||||
https://user-images.githubusercontent.com/271616/225014776-1d567049-ad71-4ef2-b050-55b0b3b9274c.mp4
|
Save the `ggml-alpaca-7b-14.bin` file in the same directory as your `./chat` executable.
|
||||||
|
|
||||||
|
The weights are based on the published fine-tunes from `alpaca-lora`, converted back into a pytorch checkpoint with a [modified script](https://github.com/tloen/alpaca-lora/pull/19) and then quantized with llama.cpp the regular way.
|
||||||
|
|
||||||
|
## Credit
|
||||||
|
|
||||||
|
This combines [Facebook's LLaMA](https://github.com/facebookresearch/llama), [Stanford Alpaca](https://crfm.stanford.edu/2023/03/13/alpaca.html), [alpaca-lora](https://github.com/tatsu-lab/stanford_alpaca) (which uses [Jason Phang's implementation of LLaMA](https://github.com/huggingface/transformers/pull/21955) on top of Hugging Face Transformers), and a modified version of [llama.cpp](https://github.com/ggerganov/llama.cpp) by Georgi Gerganov. The chat implementation is based on Matvey Soloviev's [Interactive Mode](https://github.com/ggerganov/llama.cpp/pull/61) for llama.cpp. Inspired by [Simon Willison's](https://til.simonwillison.net/llms/llama-7b-m2) getting started guide for LLaMA.
|
||||||
|
|
||||||
|
|
||||||
## Limitations
|
## Disclaimer
|
||||||
|
|
||||||
- We don't know yet how much the quantization affects the quality of the generated text
|
Note that the model weights are only to be used for research purposes, as they are derivative of LLaMA, and uses the published instruction data from the Stanford Alpaca project which is generated by OpenAI, which itself disallows the usage of its outputs to train competing models.
|
||||||
- Probably the token sampling can be improved
|
|
||||||
- The Accelerate framework is actually currently unused since I found that for tensor shapes typical for the Decoder,
|
|
||||||
there is no benefit compared to the ARM_NEON intrinsics implementation. Of course, it's possible that I simply don't
|
|
||||||
know how to utilize it properly. But in any case, you can even disable it with `LLAMA_NO_ACCELERATE=1 make` and the
|
|
||||||
performance will be the same, since no BLAS calls are invoked by the current implementation
|
|
||||||
|
|
||||||
### Contributing
|
|
||||||
|
|
||||||
- Contributors can open PRs
|
|
||||||
- Collaborators can push to branches in the `llama.cpp` repo
|
|
||||||
- Collaborators will be invited based on contributions
|
|
||||||
|
|
||||||
### Coding guidelines
|
|
||||||
|
|
||||||
- Avoid adding third-party dependencies, extra files, extra headers, etc.
|
|
||||||
- Always consider cross-compatibility with other operating systems and architectures
|
|
||||||
- Avoid fancy looking modern STL constructs, use basic `for` loops, avoid templates, keep it simple
|
|
||||||
- There are no strict rules for the code style, but try to follow the patterns in the code (indentation, spaces, etc.). Vertical alignment makes things more readable and easier to batch edit
|
|
||||||
- Clean-up any trailing whitespaces, use 4 spaces indentation, brackets on same line, `void * ptr`, `int & a`
|
|
||||||
- See [good first issues](https://github.com/ggerganov/llama.cpp/issues?q=is%3Aissue+is%3Aopen+label%3A%22good+first+issue%22) for tasks suitable for first contributions
|
|
||||||
|
|
||||||
### Misc
|
|
||||||
|
|
||||||
- Practice your C++ typing skills: https://typing-battles.ggerganov.com
|
|
||||||
|
|
|
||||||
|
|
@ -129,16 +129,16 @@ bool llama_model_load(const std::string & fname, llama_model & model, gpt_vocab
|
||||||
n_ff = ((2*(4*hparams.n_embd)/3 + hparams.n_mult - 1)/hparams.n_mult)*hparams.n_mult;
|
n_ff = ((2*(4*hparams.n_embd)/3 + hparams.n_mult - 1)/hparams.n_mult)*hparams.n_mult;
|
||||||
n_parts = LLAMA_N_PARTS.at(hparams.n_embd);
|
n_parts = LLAMA_N_PARTS.at(hparams.n_embd);
|
||||||
|
|
||||||
fprintf(stderr, "%s: n_vocab = %d\n", __func__, hparams.n_vocab);
|
// fprintf(stderr, "%s: n_vocab = %d\n", __func__, hparams.n_vocab);
|
||||||
fprintf(stderr, "%s: n_ctx = %d\n", __func__, hparams.n_ctx);
|
// fprintf(stderr, "%s: n_ctx = %d\n", __func__, hparams.n_ctx);
|
||||||
fprintf(stderr, "%s: n_embd = %d\n", __func__, hparams.n_embd);
|
// fprintf(stderr, "%s: n_embd = %d\n", __func__, hparams.n_embd);
|
||||||
fprintf(stderr, "%s: n_mult = %d\n", __func__, hparams.n_mult);
|
// fprintf(stderr, "%s: n_mult = %d\n", __func__, hparams.n_mult);
|
||||||
fprintf(stderr, "%s: n_head = %d\n", __func__, hparams.n_head);
|
// fprintf(stderr, "%s: n_head = %d\n", __func__, hparams.n_head);
|
||||||
fprintf(stderr, "%s: n_layer = %d\n", __func__, hparams.n_layer);
|
// fprintf(stderr, "%s: n_layer = %d\n", __func__, hparams.n_layer);
|
||||||
fprintf(stderr, "%s: n_rot = %d\n", __func__, hparams.n_rot);
|
// fprintf(stderr, "%s: n_rot = %d\n", __func__, hparams.n_rot);
|
||||||
fprintf(stderr, "%s: f16 = %d\n", __func__, hparams.f16);
|
// fprintf(stderr, "%s: f16 = %d\n", __func__, hparams.f16);
|
||||||
fprintf(stderr, "%s: n_ff = %d\n", __func__, n_ff);
|
// fprintf(stderr, "%s: n_ff = %d\n", __func__, n_ff);
|
||||||
fprintf(stderr, "%s: n_parts = %d\n", __func__, n_parts);
|
// fprintf(stderr, "%s: n_parts = %d\n", __func__, n_parts);
|
||||||
}
|
}
|
||||||
|
|
||||||
// load vocab
|
// load vocab
|
||||||
|
|
@ -795,7 +795,13 @@ int main(int argc, char ** argv) {
|
||||||
const int64_t t_main_start_us = ggml_time_us();
|
const int64_t t_main_start_us = ggml_time_us();
|
||||||
|
|
||||||
gpt_params params;
|
gpt_params params;
|
||||||
params.model = "models/llama-7B/ggml-model.bin";
|
|
||||||
|
params.temp = 0.1f;
|
||||||
|
params.top_p = 0.95f;
|
||||||
|
params.interactive = true;
|
||||||
|
params.interactive_start = true;
|
||||||
|
params.use_color = true;
|
||||||
|
params.model = "ggml-alpaca-7b-q4.bin";
|
||||||
|
|
||||||
if (gpt_params_parse(argc, argv, params) == false) {
|
if (gpt_params_parse(argc, argv, params) == false) {
|
||||||
return 1;
|
return 1;
|
||||||
|
|
@ -846,20 +852,28 @@ int main(int argc, char ** argv) {
|
||||||
std::vector<float> logits;
|
std::vector<float> logits;
|
||||||
|
|
||||||
// tokenize the prompt
|
// tokenize the prompt
|
||||||
std::vector<gpt_vocab::id> embd_inp = ::llama_tokenize(vocab, params.prompt, true);
|
std::vector<gpt_vocab::id> embd_inp;// = ::llama_tokenize(vocab, params.prompt, true);
|
||||||
|
|
||||||
params.n_predict = std::min(params.n_predict, model.hparams.n_ctx - (int) embd_inp.size());
|
// params.n_predict = std::min(params.n_predict, model.hparams.n_ctx - (int) embd_inp.size());
|
||||||
|
|
||||||
|
// // tokenize the reverse prompt
|
||||||
|
// std::vector<gpt_vocab::id> antiprompt_inp = ::llama_tokenize(vocab, params.antiprompt, false);
|
||||||
|
|
||||||
|
// fprintf(stderr, "\n");
|
||||||
|
// fprintf(stderr, "%s: prompt: '%s'\n", __func__, params.prompt.c_str());
|
||||||
|
// fprintf(stderr, "%s: number of tokens in prompt = %zu\n", __func__, embd_inp.size());
|
||||||
|
// for (int i = 0; i < (int) embd_inp.size(); i++) {
|
||||||
|
// fprintf(stderr, "%6d -> '%s'\n", embd_inp[i], vocab.id_to_token.at(embd_inp[i]).c_str());
|
||||||
|
// }
|
||||||
|
// fprintf(stderr, "\n");
|
||||||
|
|
||||||
|
std::vector<gpt_vocab::id> instruct_inp = ::llama_tokenize(vocab, " Below is an instruction that describes a task. Write a response that appropriately completes the request.\n\n", true);
|
||||||
|
std::vector<gpt_vocab::id> prompt_inp = ::llama_tokenize(vocab, "### Instruction:\n\n", true);
|
||||||
|
std::vector<gpt_vocab::id> response_inp = ::llama_tokenize(vocab, "### Response:\n\n", false);
|
||||||
|
|
||||||
|
embd_inp.insert(embd_inp.end(), instruct_inp.begin(), instruct_inp.end());
|
||||||
|
|
||||||
// tokenize the reverse prompt
|
|
||||||
std::vector<gpt_vocab::id> antiprompt_inp = ::llama_tokenize(vocab, params.antiprompt, false);
|
|
||||||
|
|
||||||
fprintf(stderr, "\n");
|
|
||||||
fprintf(stderr, "%s: prompt: '%s'\n", __func__, params.prompt.c_str());
|
|
||||||
fprintf(stderr, "%s: number of tokens in prompt = %zu\n", __func__, embd_inp.size());
|
|
||||||
for (int i = 0; i < (int) embd_inp.size(); i++) {
|
|
||||||
fprintf(stderr, "%6d -> '%s'\n", embd_inp[i], vocab.id_to_token.at(embd_inp[i]).c_str());
|
|
||||||
}
|
|
||||||
fprintf(stderr, "\n");
|
|
||||||
if (params.interactive) {
|
if (params.interactive) {
|
||||||
#if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__))
|
#if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__))
|
||||||
struct sigaction sigint_action;
|
struct sigaction sigint_action;
|
||||||
|
|
@ -873,14 +887,14 @@ int main(int argc, char ** argv) {
|
||||||
|
|
||||||
fprintf(stderr, "%s: interactive mode on.\n", __func__);
|
fprintf(stderr, "%s: interactive mode on.\n", __func__);
|
||||||
|
|
||||||
if(antiprompt_inp.size()) {
|
// if(antiprompt_inp.size()) {
|
||||||
fprintf(stderr, "%s: reverse prompt: '%s'\n", __func__, params.antiprompt.c_str());
|
// fprintf(stderr, "%s: reverse prompt: '%s'\n", __func__, params.antiprompt.c_str());
|
||||||
fprintf(stderr, "%s: number of tokens in reverse prompt = %zu\n", __func__, antiprompt_inp.size());
|
// fprintf(stderr, "%s: number of tokens in reverse prompt = %zu\n", __func__, antiprompt_inp.size());
|
||||||
for (int i = 0; i < (int) antiprompt_inp.size(); i++) {
|
// for (int i = 0; i < (int) antiprompt_inp.size(); i++) {
|
||||||
fprintf(stderr, "%6d -> '%s'\n", antiprompt_inp[i], vocab.id_to_token.at(antiprompt_inp[i]).c_str());
|
// fprintf(stderr, "%6d -> '%s'\n", antiprompt_inp[i], vocab.id_to_token.at(antiprompt_inp[i]).c_str());
|
||||||
}
|
// }
|
||||||
fprintf(stderr, "\n");
|
// fprintf(stderr, "\n");
|
||||||
}
|
// }
|
||||||
}
|
}
|
||||||
fprintf(stderr, "sampling parameters: temp = %f, top_k = %d, top_p = %f, repeat_last_n = %i, repeat_penalty = %f\n", params.temp, params.top_k, params.top_p, params.repeat_last_n, params.repeat_penalty);
|
fprintf(stderr, "sampling parameters: temp = %f, top_k = %d, top_p = %f, repeat_last_n = %i, repeat_penalty = %f\n", params.temp, params.top_k, params.top_p, params.repeat_last_n, params.repeat_penalty);
|
||||||
fprintf(stderr, "\n\n");
|
fprintf(stderr, "\n\n");
|
||||||
|
|
@ -897,7 +911,7 @@ int main(int argc, char ** argv) {
|
||||||
|
|
||||||
|
|
||||||
if (params.interactive) {
|
if (params.interactive) {
|
||||||
fprintf(stderr, "== Running in interactive mode. ==\n"
|
fprintf(stderr, "== Running in chat mode. ==\n"
|
||||||
#if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__)) || defined (_WIN32)
|
#if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__)) || defined (_WIN32)
|
||||||
" - Press Ctrl+C to interject at any time.\n"
|
" - Press Ctrl+C to interject at any time.\n"
|
||||||
#endif
|
#endif
|
||||||
|
|
@ -907,7 +921,7 @@ int main(int argc, char ** argv) {
|
||||||
|
|
||||||
int remaining_tokens = params.n_predict;
|
int remaining_tokens = params.n_predict;
|
||||||
int input_consumed = 0;
|
int input_consumed = 0;
|
||||||
bool input_noecho = false;
|
bool input_noecho = true;
|
||||||
|
|
||||||
// prompt user immediately after the starting prompt has been loaded
|
// prompt user immediately after the starting prompt has been loaded
|
||||||
if (params.interactive_start) {
|
if (params.interactive_start) {
|
||||||
|
|
@ -919,7 +933,9 @@ int main(int argc, char ** argv) {
|
||||||
printf(ANSI_COLOR_YELLOW);
|
printf(ANSI_COLOR_YELLOW);
|
||||||
}
|
}
|
||||||
|
|
||||||
while (remaining_tokens > 0) {
|
|
||||||
|
|
||||||
|
while (true) {
|
||||||
// predict
|
// predict
|
||||||
if (embd.size() > 0) {
|
if (embd.size() > 0) {
|
||||||
const int64_t t_start_us = ggml_time_us();
|
const int64_t t_start_us = ggml_time_us();
|
||||||
|
|
@ -935,7 +951,7 @@ int main(int argc, char ** argv) {
|
||||||
n_past += embd.size();
|
n_past += embd.size();
|
||||||
embd.clear();
|
embd.clear();
|
||||||
|
|
||||||
if (embd_inp.size() <= input_consumed) {
|
if (embd_inp.size() <= input_consumed && !is_interacting) {
|
||||||
// out of user input, sample next token
|
// out of user input, sample next token
|
||||||
const float top_k = params.top_k;
|
const float top_k = params.top_k;
|
||||||
const float top_p = params.top_p;
|
const float top_p = params.top_p;
|
||||||
|
|
@ -964,10 +980,12 @@ int main(int argc, char ** argv) {
|
||||||
input_noecho = false;
|
input_noecho = false;
|
||||||
|
|
||||||
// decrement remaining sampling budget
|
// decrement remaining sampling budget
|
||||||
--remaining_tokens;
|
// --remaining_tokens;
|
||||||
} else {
|
} else {
|
||||||
// some user input remains from prompt or interaction, forward it to processing
|
// some user input remains from prompt or interaction, forward it to processing
|
||||||
while (embd_inp.size() > input_consumed) {
|
while (embd_inp.size() > input_consumed) {
|
||||||
|
// fprintf(stderr, "%6d -> '%s'\n", embd_inp[input_consumed], vocab.id_to_token.at(embd_inp[input_consumed]).c_str());
|
||||||
|
|
||||||
embd.push_back(embd_inp[input_consumed]);
|
embd.push_back(embd_inp[input_consumed]);
|
||||||
last_n_tokens.erase(last_n_tokens.begin());
|
last_n_tokens.erase(last_n_tokens.begin());
|
||||||
last_n_tokens.push_back(embd_inp[input_consumed]);
|
last_n_tokens.push_back(embd_inp[input_consumed]);
|
||||||
|
|
@ -995,11 +1013,19 @@ int main(int argc, char ** argv) {
|
||||||
// check if we should prompt the user for more
|
// check if we should prompt the user for more
|
||||||
if (params.interactive && embd_inp.size() <= input_consumed) {
|
if (params.interactive && embd_inp.size() <= input_consumed) {
|
||||||
// check for reverse prompt
|
// check for reverse prompt
|
||||||
if (antiprompt_inp.size() && std::equal(antiprompt_inp.rbegin(), antiprompt_inp.rend(), last_n_tokens.rbegin())) {
|
// if (antiprompt_inp.size() && std::equal(antiprompt_inp.rbegin(), antiprompt_inp.rend(), last_n_tokens.rbegin())) {
|
||||||
// reverse prompt found
|
// // reverse prompt found
|
||||||
is_interacting = true;
|
// is_interacting = true;
|
||||||
}
|
// }
|
||||||
if (is_interacting) {
|
if (is_interacting) {
|
||||||
|
// input_consumed = 0;
|
||||||
|
// embd_inp.erase(embd_inp.begin());
|
||||||
|
input_consumed = embd_inp.size();
|
||||||
|
embd_inp.insert(embd_inp.end(), prompt_inp.begin(), prompt_inp.end());
|
||||||
|
|
||||||
|
|
||||||
|
printf("\n> ");
|
||||||
|
|
||||||
// currently being interactive
|
// currently being interactive
|
||||||
bool another_line=true;
|
bool another_line=true;
|
||||||
while (another_line) {
|
while (another_line) {
|
||||||
|
|
@ -1026,8 +1052,7 @@ int main(int argc, char ** argv) {
|
||||||
|
|
||||||
std::vector<gpt_vocab::id> line_inp = ::llama_tokenize(vocab, buf, false);
|
std::vector<gpt_vocab::id> line_inp = ::llama_tokenize(vocab, buf, false);
|
||||||
embd_inp.insert(embd_inp.end(), line_inp.begin(), line_inp.end());
|
embd_inp.insert(embd_inp.end(), line_inp.begin(), line_inp.end());
|
||||||
|
embd_inp.insert(embd_inp.end(), response_inp.begin(), response_inp.end());
|
||||||
remaining_tokens -= line_inp.size();
|
|
||||||
|
|
||||||
input_noecho = true; // do not echo this again
|
input_noecho = true; // do not echo this again
|
||||||
}
|
}
|
||||||
|
|
@ -1038,8 +1063,9 @@ int main(int argc, char ** argv) {
|
||||||
|
|
||||||
// end of text token
|
// end of text token
|
||||||
if (embd.back() == 2) {
|
if (embd.back() == 2) {
|
||||||
fprintf(stderr, " [end of text]\n");
|
// fprintf(stderr, " [end of text]\n");
|
||||||
break;
|
is_interacting = true;
|
||||||
|
continue;
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
BIN
screencast.gif
Normal file
BIN
screencast.gif
Normal file
Binary file not shown.
|
After Width: | Height: | Size: 1.2 MiB |
Loading…
Add table
Add a link
Reference in a new issue