fgerlits commented on code in PR #1903: URL: https://github.com/apache/nifi-minifi-cpp/pull/1903#discussion_r2039102001
########## extensions/llamacpp/processors/RunLlamaCppInference.cpp: ########## @@ -0,0 +1,164 @@ +/** + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include "RunLlamaCppInference.h" +#include "core/ProcessContext.h" +#include "core/ProcessSession.h" +#include "core/Resource.h" +#include "Exception.h" + +#include "rapidjson/document.h" +#include "rapidjson/error/en.h" +#include "LlamaContext.h" +#include "utils/ProcessorConfigUtils.h" + +namespace org::apache::nifi::minifi::extensions::llamacpp::processors { + +void RunLlamaCppInference::initialize() { + setSupportedProperties(Properties); + setSupportedRelationships(Relationships); +} + +void RunLlamaCppInference::onSchedule(core::ProcessContext& context, core::ProcessSessionFactory&) { + model_path_.clear(); + model_path_ = utils::parseProperty(context, ModelPath); + system_prompt_ = context.getProperty(SystemPrompt).value_or(""); + + LlamaSamplerParams llama_sampler_params; + llama_sampler_params.temperature = utils::parseOptionalFloatProperty(context, Temperature); + if (auto top_k = utils::parseOptionalI64Property(context, TopK)) { + llama_sampler_params.top_k = gsl::narrow<int32_t>(*top_k); + } + llama_sampler_params.top_p = utils::parseOptionalFloatProperty(context, TopP); + llama_sampler_params.min_p = utils::parseOptionalFloatProperty(context, MinP); + llama_sampler_params.min_keep = utils::parseU64Property(context, MinKeep); + + LlamaContextParams llama_ctx_params; + llama_ctx_params.n_ctx = gsl::narrow<uint32_t>(utils::parseU64Property(context, TextContextSize)); + llama_ctx_params.n_batch = gsl::narrow<uint32_t>(utils::parseU64Property(context, LogicalMaximumBatchSize)); + llama_ctx_params.n_ubatch = gsl::narrow<uint32_t>(utils::parseU64Property(context, PhysicalMaximumBatchSize)); + llama_ctx_params.n_seq_max = gsl::narrow<uint32_t>(utils::parseU64Property(context, MaxNumberOfSequences)); + llama_ctx_params.n_threads = gsl::narrow<int32_t>(utils::parseI64Property(context, ThreadsForGeneration)); + llama_ctx_params.n_threads_batch = gsl::narrow<int32_t>(utils::parseI64Property(context, ThreadsForBatchProcessing)); + + llama_ctx_ = LlamaContext::create(model_path_, llama_sampler_params, llama_ctx_params); +} + +void RunLlamaCppInference::increaseTokensIn(uint64_t token_count) { + auto* const llamacpp_metrics = dynamic_cast<RunLlamaCppInferenceMetrics*>(metrics_.get()); + gsl_Assert(llamacpp_metrics); + std::lock_guard<std::mutex> lock(llamacpp_metrics->tokens_in_mutex_); + if (llamacpp_metrics->tokens_in > std::numeric_limits<uint64_t>::max() - token_count) { + logger_->log_warn("Tokens in count overflow detected, resetting to 0"); + llamacpp_metrics->tokens_in = token_count; + return; + } + + llamacpp_metrics->tokens_in += token_count; +} + +void RunLlamaCppInference::increaseTokensOut(uint64_t token_count) { + auto* const llamacpp_metrics = dynamic_cast<RunLlamaCppInferenceMetrics*>(metrics_.get()); + gsl_Assert(llamacpp_metrics); + std::lock_guard<std::mutex> lock(llamacpp_metrics->tokens_out_mutex_); + if (llamacpp_metrics->tokens_out > std::numeric_limits<uint64_t>::max() - token_count) { + logger_->log_warn("Tokens out count overflow detected, resetting to 0"); + llamacpp_metrics->tokens_out = token_count; + return; + } + + llamacpp_metrics->tokens_out += token_count; +} + +void RunLlamaCppInference::onTrigger(core::ProcessContext& context, core::ProcessSession& session) { + auto flow_file = session.get(); + if (!flow_file) { + context.yield(); + return; + } + + auto prompt = context.getProperty(Prompt, flow_file.get()).value_or(""); + + auto read_result = session.readBuffer(flow_file); + std::string input_data_and_prompt; + if (!read_result.buffer.empty()) { + input_data_and_prompt.append("Input data (or flow file content):\n"); + input_data_and_prompt.append({reinterpret_cast<const char*>(read_result.buffer.data()), read_result.buffer.size()}); + input_data_and_prompt.append("\n\n"); + } + input_data_and_prompt.append(prompt); + + if (input_data_and_prompt.empty()) { + logger_->log_error("Input data and prompt are empty"); + session.transfer(flow_file, Failure); + return; + } + + auto input = [&] { + std::vector<LlamaChatMessage> messages; + if (!system_prompt_.empty()) { + messages.push_back({.role = "system", .content = system_prompt_}); + } + messages.push_back({.role = "user", .content = input_data_and_prompt}); + + return llama_ctx_->applyTemplate(messages); + }(); + + if (!input) { + logger_->log_error("Inference failed with while applying template"); + session.transfer(flow_file, Failure); + return; + } + + logger_->log_debug("AI model input: {}", *input); + + auto start_time = std::chrono::steady_clock::now(); + + std::string text; + auto generation_result = llama_ctx_->generate(*input, [&] (std::string_view token) { + text += token; + }); + + auto elapsed_time = std::chrono::duration_cast<std::chrono::milliseconds>(std::chrono::steady_clock::now() - start_time).count(); + + if (!generation_result) { + logger_->log_error("Inference failed with generation error: '{}'", generation_result.error()); + session.transfer(flow_file, Failure); + return; + } + + increaseTokensIn(generation_result->num_tokens_in); + increaseTokensOut(generation_result->num_tokens_out); + + logger_->log_debug("Number of tokens generated: {}", generation_result->num_tokens_out); + logger_->log_debug("AI model inference time: {} ms", elapsed_time); + logger_->log_debug("AI model output: {}", text); + + flow_file->setAttribute(LlamaCppTimeToFirstToken.name, std::to_string(generation_result->time_to_first_token.count()) + " ms"); + flow_file->setAttribute(LlamaCppTokensPerSecond.name, std::to_string(generation_result->tokens_per_second)); Review Comment: `to_string` prints 6 digits after the decimal point; I think 2 would be enough: ```suggestion flow_file->setAttribute(LlamaCppTokensPerSecond.name, fmt::format("{:.2}", generation_result->tokens_per_second)); ``` edit: or maybe even just {:.1}? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@nifi.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
