FiveTech Software tech support forums

... model.save_pretrained_gguf( "folderName", tokenizer, quantization_method = "f16")

Modelfile
FROM ./unsloth.BF16.gguf
PARAMETER temperature 0.7
PARAMETER num_ctx 1024
SYSTEM """You are Mario from super mario bros acting as an assistant"""

ollama create myModel --file Modelfile

ollama run myModel

... Requiere llama.cpp instalado
python3 llama.cpp/convert-hf-to-gguf.py merged_model/ \
--outtype f16 \
--outfile qwen2.5-3b-instruct-merged.gguf

Modelfile
FROM ./qwen2.5-3b-instruct-merged.gguf
PARAMETER temperature 0.8
PARAMETER num_ctx 1024

ollama create my-model -f Modelfile
ollama run my-model

... editing tool，open and modify the system prompt and template in the model file to suit your preferences or requirements:

PARAMETER temperature 1
PARAMETER num_ctx 4096
SYSTEM """You are Mario from super mario bros acting as an assistant"""

ollama create myllama2 --file myllama2.modelfile

ollama list

... rank,
gpu_memory_utilization = 0.5, # Reduce if out of memory
)

model = FastLanguageModel.get_peft_model(
model,
r = lora_rank, # Choose any number > 0 ! Suggested 8, 16, 32, 64, 128
target_modules = [
"q_proj", "k_proj", "v_proj", "o_proj",
"gate_proj", "up_proj", "down_proj ...

... Qwen/Qwen2.5-3B-Instruct")

# gemerate r1 prompt with a prefix for the model to already start with the thinking process
def generate_r1_prompt(numbers, target):
r1_prefix = [{
"role": "system",
"content": "You are a helpful assistant. You first thinks about the reasoning process in the mind ...

Ollama defaults to 2k context window. To increase it:

ollama run deepseek-r1:32b
/set parameter num_ctx 8192
/save deepseek-r1:32b-8k
ollama serve

... SIZE 195,10
@ 16.8, 35 SAY "Set Century is on." OF oDlg SIZE 195,10
@ 17.3, 35 SAY "Set Epoch is set to 1919." OF oDlg SIZE 195,10
@ 18, 03 SAY " Num PICT '99' !>10 10" OF oDlg COLORS CLR_BLACK ,CLR_HCYAN BORDER SIZE 195,10
@ 19, 03 SAY " Memo No PICT 256" OF oDlg COLORS CLR_BLACK ,CLR_HCYAN ...

#include "FiveWin.ch"

// Función principal para entrenar el Transformer
PROCEDURE Main()
LOCAL d_model := 128
LOCAL n_heads := 4
LOCAL num_layers := 1
LOCAL learning_rate := 0.001
LOCAL max_epochs := 100
LOCAL output, loss, d_output, epoch

// Generar datos de entrada y salida aleatorios ...

#include "FiveWin.ch"

// Función principal para entrenar el Transformer
PROCEDURE Main()
LOCAL d_model := 128
LOCAL n_heads := 4
LOCAL num_layers := 1
LOCAL learning_rate := 0.01
LOCAL max_epochs := 100
LOCAL output, loss, d_output, epoch

// Generar datos de entrada y salida aleatorios ...

... return hb_MatrixZero(::d_model, ::d_model) // Simplificación para no propagar más

// Clase para el Transformer
CLASS Transformer
DATA layers
DATA num_layers

METHOD New(num_layers, d_model, n_heads)
METHOD Forward(src, tgt)
METHOD Backward(d_output)
ENDCLASS

METHOD New(num_layers, d_model, n ...

... egreso), s1 = ( @nT1 := saldo ), ingreso_val = (ingreso * costo ), saldo_val = ( @nT2 + ingreso_val - egreso_val), s2 = (@nT2 := saldo_val ), kardex.num_movimiento = (@nT3 := @nT3 + 1), s1 = IF(s1 <= 0, 1, s1), precio_promedio = IF(s2 / s1 >= 0, s2 / s1,0), costo = IF(egreso > 0, If(precio_promedio ...

... 0, 0, 0 } }

XBROWSER aData SETUP TITLE "SetColsAsRows" xSetUp( oBrw )

return nil

function xSetUp( oBrw )

local n

oBrw:cHeaders := { "Name", "Num/Rit", "Rit", "Num/Rit", "Rit" }
AEval( oBrw:aCols, { |oCol,i| If( i % 2 = 0, ;
oCol:bClrStd := oCol:bClrSelFocus := { || { CLR_WHITE, CLR_HRED ...

I resolved with



oPrn:Ellipse(...)

Hi Marco
not only how many tags but its formula for that single tag
e.g.: dtos( gio_drg) + str(gio_num,9) + str(gio_riga,3)
I often made too complex indexes which often slowed down the browse,
I would like to know if the slowness you see is on loop type : ' do while .. enddo '
or on a browse
Bye

Estimado Esteban,

Buscando el primer símbolo que no te encuentra:
hbmysql.lib(mysql.obj) : error LNK2019: s¡mbolo externo server_mysql_num_rows sin resolver al que se hace referencia en la funci¢n HB_FUN_MYSQL_LIST_DBS

server_mysql_num_rows no aparece en ningún fichero de los que hay en https ...