input too long skipped 6 tokens.

[iplayfast]

I don't understand how this input is too long. This is what happened.

./main -t 30 -ngl 40 -m AllModels/wizardcoder-python-34b-v1.0.Q5_K_M.gguf --color -c 32K --temp 0.7 --repeat_penalty 1.1 -n -1 -ins -p "write out the steps needed to create a snake game in python"
Log start
main: build = 1176 (bd33e5a)
main: seed  = 1693849555
ggml_init_cublas: found 1 CUDA devices:
  Device 0: NVIDIA GeForce RTX 4090, compute capability 8.9
llama_model_loader: loaded meta data with 17 key-value pairs and 435 tensors from AllModels/wizardcoder-python-34b-v1.0.Q5_K_M.gguf (version GGUF V1 (support until nov 2023))
llama_model_loader: - tensor    0:                token_embd.weight q4_0     [  8192, 32001,     1,     1 ]
llama_model_loader: - tensor    1:              blk.0.attn_q.weight q5_K     [  8192,  8192,     1,     1 ]
llama_model_loader: - tensor    2:              blk.0.attn_k.weight q5_K     [  8192,  1024,     1,     1 ]
llama_model_loader: - tensor    3:              blk.0.attn_v.weight q6_K     [  8192,  1024,     1,     1 ]
llama_model_loader: - tensor    4:         blk.0.attn_output.weight q5_K     [  8192,  8192,     1,     1 ]
llama_model_loader: - tensor    5:            blk.0.ffn_gate.weight q5_K     [  8192, 22016,     1,     1 ]
llama_model_loader: - tensor    6:              blk.0.ffn_up.weight q5_K     [  8192, 22016,     1,     1 ]
llama_model_loader: - tensor    7:            blk.0.ffn_down.weight q6_K     [ 22016,  8192,     1,     1 ]
llama_model_loader: - tensor    8:           blk.0.attn_norm.weight f32      [  8192,     1,     1,     1 ]

...

llama_model_loader: - type q6_K:   48 tensors
llm_load_print_meta: format         = GGUF V1 (support until nov 2023)
llm_load_print_meta: arch           = llama
llm_load_print_meta: vocab type     = SPM
llm_load_print_meta: n_vocab        = 32001
llm_load_print_meta: n_merges       = 0
llm_load_print_meta: n_ctx_train    = 16384
llm_load_print_meta: n_ctx          = 32
llm_load_print_meta: n_embd         = 8192
llm_load_print_meta: n_head         = 64
llm_load_print_meta: n_head_kv      = 8
llm_load_print_meta: n_layer        = 48
llm_load_print_meta: n_rot          = 128
llm_load_print_meta: n_gqa          = 8
llm_load_print_meta: f_norm_eps     = 1.0e-05
llm_load_print_meta: f_norm_rms_eps = 1.0e-05
llm_load_print_meta: n_ff           = 22016
llm_load_print_meta: freq_base      = 1000000.0
llm_load_print_meta: freq_scale     = 1
llm_load_print_meta: model type     = 34B
llm_load_print_meta: model ftype    = mostly Q5_K - Medium
llm_load_print_meta: model size     = 33.74 B
llm_load_print_meta: general.name   = LLaMA
llm_load_print_meta: BOS token = 1 '<s>'
llm_load_print_meta: EOS token = 2 '</s>'
llm_load_print_meta: UNK token = 0 '<unk>'
llm_load_print_meta: LF token  = 13 '<0x0A>'
llm_load_tensors: ggml ctx size =    0.14 MB
llm_load_tensors: using CUDA for GPU acceleration
llm_load_tensors: mem required  = 4414.75 MB (+    6.00 MB per state)
llm_load_tensors: offloading 40 repeating layers to GPU
llm_load_tensors: offloaded 40/51 layers to GPU
llm_load_tensors: VRAM used: 18583 MB
..................................................................................................
llama_new_context_with_model: kv self size  =    6.00 MB
llama_new_context_with_model: compute buffer total size =    8.84 MB
llama_new_context_with_model: VRAM scratch buffer: 7.38 MB

system_info: n_threads = 30 / 32 | AVX = 1 | AVX2 = 1 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | FMA = 1 | NEON = 0 | ARM_FMA = 0 | F16C = 1 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | 
main: interactive mode on.
Reverse prompt: '### Instruction:

'
sampling: repeat_last_n = 64, repeat_penalty = 1.100000, presence_penalty = 0.000000, frequency_penalty = 0.000000, top_k = 40, tfs_z = 1.000000, top_p = 0.950000, typical_p = 1.000000, temp = 0.700000, mirostat = 0, mirostat_lr = 0.100000, mirostat_ent = 5.000000
generate: n_ctx = 32, n_batch = 512, n_predict = -1, n_keep = 14


== Running in interactive mode. ==
 - Press Ctrl+C to interject at any time.
 - Press Return to return control to LLaMa.
 - To return control without starting a new line, end your input with '/'.
 - If you want to submit another line, end your input with '\'.

 write out the steps needed to create a snake game in python
> write out the steps needed to create a snake game in python
<<input too long: skipped 6 tokens>>GGML_ASSERT: ggml.c:4762: view_src == NULL || data_size + view_offs <= ggml_nbytechricchchchcccchris@FORGE:~/ai/llama.cpp$ ./main -t 30 -ngl 40 -m AllModels/wizardcoder-python-34b-v1.0.Q5_K_M.gguf --color -c 32K --temp 0.7 --repeat_penalty 1.1 -n -1 -ins -p "write out the steps needed to create a snake game in python"

[KerfuffleV2]

-c 32K doesn't work the way you expect. The parsing of integer arguments is pretty simple, it just parses what it can. So it parses 32, hits K and stops. You end up with a context size of 32. Use -c 32000 if you want 32,000.

[iplayfast]

forehead slap! Thanks!