sigmas as std::string instead of vector

parsing issue : more convenient as [,,,]

Author: piallai

examples/cli/main_glove.h

@@ -195,7 +195,7 @@ glvm_parametrization(GlvSdGenerationOptionsAdvanced, "Generation options advance
     clip_skip, int, "--clip-skip", "ignore last layers of CLIP network; 1 ignores none, 2 ignores one layer (default: -1) \n<= 0 represents unspecified, will be 1 for SD1.x, 2 for SD2.x", -1,
     moe_boundary, float, "--moe-boundary", "timestep boundary for Wan2.2 MoE model. (default: 0.875). Only enabled if `--high-noise-steps` is set to -1", 0.875f,
     vace_strength, float, "--vace-strength", "wan vace strength", 1.f,
-    sigmas, std::vector<float>, "--sigmas", "custom sigma values for the sampler, comma-separated (e.g., \"14.61,7.8,3.5,0.0\").\nCan not set values for now. Parsing of values would be more convenient in a vector format such as: [14.61,7.8,3.5,0.0]. Alike skip layers.", {},    
+    sigmas, std::string, "--sigmas", "custom sigma values for the sampler, comma-separated (e.g., \"14.61,7.8,3.5,0.0\").", {},    
     increase_ref_index, bool, "--increase-ref-index", "automatically increase the indices of references images based on the order they are listed (starting with 1).", false,
     sampling_method, SamplingMethod, "--sampling-method", "{euler, euler_a, heun, dpm2, dpm++2s_a, dpm++2m, dpm++2mv2, ipndm, ipndm_v, lcm, ddim_trailing, tcd} \nsampling method (default: 'euler_a')", SamplingMethod::euler_a,
     scheduler, Scheduler, "--scheduler", "denoiser sigma scheduler, one of [discrete, karras, exponential, ays, gits, smoothstep, sgm_uniform, simple, kl_optimal, lcm],\ndefault: discrete", Scheduler::discrete,