mirror of
https://github.com/ggerganov/llama.cpp.git
synced 2024-12-26 14:20:31 +01:00
main : option to disable context shift (#9484)
* added cli arg to disable context shift * reverted precommit * updated README.md for main * white space * allow disabling context shift in the server * Update common/arg.cpp no-context-shift only works for main example Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> * added server example to --no-context-shift args * removed server changes * white space --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit is contained in:
parent
c4965a64f7
commit
441b72b91f
@ -685,6 +685,13 @@ gpt_params_context gpt_params_parser_init(gpt_params & params, llama_example ex,
|
||||
params.n_keep = value;
|
||||
}
|
||||
));
|
||||
add_opt(llama_arg(
|
||||
{"--no-context-shift"},
|
||||
format("disables context shift on inifinite text generation (default: %s)", params.ctx_shift ? "disabled" : "enabled"),
|
||||
[](gpt_params & params) {
|
||||
params.ctx_shift = false;
|
||||
}
|
||||
).set_examples({LLAMA_EXAMPLE_MAIN}));
|
||||
add_opt(llama_arg(
|
||||
{"--chunks"}, "N",
|
||||
format("max number of chunks to process (default: %d, -1 = all)", params.n_chunks),
|
||||
@ -1985,4 +1992,3 @@ gpt_params_context gpt_params_parser_init(gpt_params & params, llama_example ex,
|
||||
|
||||
return ctx_arg;
|
||||
}
|
||||
|
||||
|
@ -246,6 +246,7 @@ struct gpt_params {
|
||||
bool cont_batching = true; // insert new sequences for decoding on-the-fly
|
||||
bool flash_attn = false; // flash attention
|
||||
bool no_perf = false; // disable performance metrics
|
||||
bool ctx_shift = true; // context shift on inifinite text generation
|
||||
|
||||
bool input_prefix_bos = false; // prefix BOS to user inputs, preceding input_prefix
|
||||
bool logits_all = false; // return logits for all tokens in the batch
|
||||
|
@ -161,6 +161,8 @@ A value of -1 will enable infinite text generation, even though we have a finite
|
||||
|
||||
If the pause is undesirable, a value of -2 will stop generation immediately when the context is filled.
|
||||
|
||||
The `--no-context-shift` option allows you to stop the infinite text generation once the finite context window is full.
|
||||
|
||||
It is important to note that the generated text may be shorter than the specified number of tokens if an End-of-Sequence (EOS) token or a reverse prompt is encountered. In interactive mode, text generation will pause and control will be returned to the user. In non-interactive mode, the program will end. In both cases, the text generation may stop before reaching the specified `--predict` value. If you want the model to keep going without ever producing End-of-Sequence on its own, you can use the `--ignore-eos` parameter.
|
||||
|
||||
### Temperature
|
||||
|
@ -559,7 +559,12 @@ int main(int argc, char ** argv) {
|
||||
// if we run out of context:
|
||||
// - take the n_keep first tokens from the original prompt (via n_past)
|
||||
// - take half of the last (n_ctx - n_keep) tokens and recompute the logits in batches
|
||||
|
||||
if (n_past + (int) embd.size() >= n_ctx) {
|
||||
if (!params.ctx_shift){
|
||||
LOG_DBG("\n\n%s: context full and context shift is disabled => stopping\n", __func__);
|
||||
break;
|
||||
} else {
|
||||
if (params.n_predict == -2) {
|
||||
LOG_DBG("\n\n%s: context full and n_predict == -%d => stopping\n", __func__, params.n_predict);
|
||||
break;
|
||||
@ -583,6 +588,7 @@ int main(int argc, char ** argv) {
|
||||
LOG_DBG("clear session path\n");
|
||||
path_session.clear();
|
||||
}
|
||||
}
|
||||
} else {
|
||||
// context extension via Self-Extend
|
||||
while (n_past >= ga_i + ga_w) {
|
||||
|
Loading…
Reference in New Issue
Block a user